ROOT FIX (não paliativo) das 3 lacunas que travavam o Construtor:
1. get_assistant_pricing_tool: lia de Captain::Mcp::PricingTables::TABLES
(hash Ruby) que NÃO EXISTE MAIS desde a migração pra DB. Caía no
fallback de scenario raw. Refactor: lê de Captain::PricingCategory +
Captain::PricingAmount, formata grid markdown agrupado por day_bucket.
2. save_agent_spec_tool: Construtor salvava REFERÊNCIAS
(pricing_source.copied_from_assistant_id) mas hermes-provision script
espera DADOS EXPANDIDOS (categories[] com amounts, soul_md+skill_md).
Refactor: tool agora EXPANDE server-side — busca PricingCategory do
parent, monta categories array, gera SOUL.md (template + identity +
disclosure_policy) e SKILL.md (template + pricing + rules + identity).
Output já é spec consumível pelo script.
3. Captain::PricingAmount::PERIODS: adicionado '1h' (Prime tem 1h).
4. Seed pras 3 units faltando: Hotel Recanto (1) + PrimeAL (2) + Qnn01
(3). Agora os 6 units existentes têm pricing em DB.
Hot-patched ambos tools + USR1 no Puma. Construtor pronto pra criar
Bianca/Juliana etc end-to-end sem intervenção manual.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Pequenos ajustes em Captain::Unit (app + enterprise), migration de seed
inicial dos prompts Jasmine/Daniela, schema regenerado, e atualização do
README de seed_prompts pra refletir o estado atual dos modelos.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Hoje as métricas e séries do BotReports agregam toda a conta — não dá pra ver
"a Jasmine da PrimeAL está errando mais que a do Qnn01". Cada unidade tem
prompt próprio, então um sintoma localizado precisa de medição localizada.
Backend:
- Inbox#has_many :reporting_events (relação inversa que faltava)
- BotMetricsBuilder aceita inbox_id e filtra bot_conversations + base_reporting_events
- bot_metrics endpoint passa inbox_id pelos params permitidos
- count_report_builder já suporta scope=inbox; agora funciona pra
bot_resolutions_count e bot_handoffs_count graças à relação acima
Frontend:
- BotReports.vue: ReportFilters com filter-type='inboxes' e dropdown ativo
- Quando uma inbox é escolhida, requestPayload inclui inboxId/type/id e os
fetches (BotMetrics + ReportContainer) passam o filtro
- API client getBotMetrics aceita inboxId; getBotSummary aceita type+id
- Sem inbox selecionada: comportamento antigo (agregação da conta)
Bonus na rake task de retroativo:
- rebuild_bot_resolved.rake: Message.unscope(:order) pra evitar conflito
PG::InvalidColumnReference (DISTINCT + ORDER BY default scope)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Banner agressivo passa de uma notificação só ("status→open") pra
um sistema de escalada baseado em inatividade quando o cliente é
o último a falar.
Níveis:
- 5 min sem resposta → AMARELO, sem som
- 15 min sem resposta → LARANJA, beep 1x + notificação do SO
- 28 min sem resposta → VERMELHO pulsante + som em loop infinito
- status→open (reabertura) → VERMELHO imediato
Por conversa, o banner mostra um item com nome do contato, inbox
e contexto ("reabriu agora" / "15 min sem resposta"). Headline
grande e explicação clara sobre como limpa.
Comportamento do × dismiss:
- Antes: apagava o alerta de vez. Agente podia "fingir que viu".
- Agora: esconde temporariamente. Volta quando escalar (próximo
threshold) ou nova mensagem. A única forma de LIMPAR de vez é
responder o cliente (tracker detecta msg outgoing do User ou
AgentBot e chama dismissForReply).
Permissões:
- account.settings.aggressive_alert_enabled (master switch admin)
- user.ui_settings.aggressive_alert_enabled (toggle do próprio agente)
- Default true pros dois; um false em qualquer bloqueia alertas.
Settings UI:
- Conta → General: novo card "Alerta agressivo (master switch)"
- Perfil do usuário: novo card "Receber alertas agressivos"
Arquivos:
- helper/aggressiveAlert.js: multi-level state, hide vs dismiss-for-reply
- helper/inactivityAlertTracker.js: timer único, thresholds declarativos
- helper/actionCable.js: hook em onMessageCreated (feed tracker) +
isAggressiveAlertEnabled() + limpa tracker em status_changed != open
- components/app/AggressiveConversationBanner.vue: variantes de cor,
headline grande, explanation, × temp-hide
- account.rb + accounts_controller.rb: store_accessor + permitted
- settings UI components (account + profile): switches auto-persist
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Resolve duas camadas de problema identificadas em teste end-to-end:
1. Embeddings falhavam com HTTP 404 (/codex/v1/embeddings não existe).
Solução: Captain::Llm::EmbeddingService sempre usa OpenAI tradicional
via Llm::Config.with_api_key(legacy_settings). ProviderConfig expõe
legacy_openai_settings pra isso.
2. Servidor Codex ocasionalmente responde com response.failed +
code=server_error (instabilidade transitória). Client agora retenta
até 2x com backoff exponencial (0.5s, 1.5s) em erros retryable:
HTTP 5xx, server_error no response.failed, ou stream inacabado.
Outras correções nesta etapa:
- Scenario#agent_model: em modo Codex, ignora CAPTAIN_OPEN_AI_MODEL_SCENARIO
(que pode ter gpt-4o legado) e usa ProviderConfig.model.
- ExtractionService/ContradictionCheckerService/TranslateQueryService:
trocam constantes hardcoded gpt-4o-mini/gpt-4.1-nano por
ProviderConfig.light_model (respeitando o provider ativo).
- ProviderConfig.DEFAULT_CODEX_MODEL agora é gpt-5.2 (reconhecido pelo
RubyLLM; gpt-5.4 não está no catalog do gem).
Validado ponta-a-ponta: WhatsApp → Chatwoot → Jasmine → handoff Daniela
→ faq_lookup com embedding OK → resposta com preços corretos.
Docs em docs/captain-codex-oauth.md.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Consolida o trabalho desta branch de abril/2026 em um bloco pronto pra
testar em staging antes do merge pra main.
## Correções de memória semântica
- ExtractionService: Princípio Zero + Regra de Ouro (ação consumada vs intenção).
- Cenário Daniela_Reservas: Passo 0 de classificação (consulta/intenção/fora).
## Roleta da Sorte (end-to-end)
- Schema Supabase + 7 RPCs atômicas (server-side, idempotentes).
- Services: Offer, Redeem, WeeklyReport.
- Jobs: OfferRouletteJob (hook em ConfirmationService após Pix pago),
NotifyRevealed + Scheduler de fallback.
- Tool manual GenerateRoletaLinkTool + endpoint público /roleta/notify.
- Dashboard /captain/roleta com Resgate + Relatório + anomaly detection.
## Cenário Reclamacoes_Ouvidoria
- Triagem P1-P4, framework LAST, Three-level listening, Self-check.
- Sem compensação material, detecção de cliente frustrado eleva prioridade.
## Analytics
- Funil de conversão /captain/funnel: 5 etapas via regex, zero LLM.
- Detector de churn via ChurnOutreach* (cron dias úteis 10h-17h BRT).
## Trabalho pré-existente incluído
- Captain Executive Reports (ceo_digest, mattermost_delivery).
- get_reserva_preco_tool, Lifecycle ajustes, Reservations UI polimentos.
## Outros
- .gitignore: patterns pra credenciais.
- Migrations de scenarios idempotentes.
- i18n completa pt_BR+en pra roleta/funnel.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Inboxes without portal_id were crashing with NoMethodError on save,
blocking landing host creation via UI for any inbox without a portal.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Cria modelo LeadClick para registrar cliques das landing pages
- Cria modelo LandingHost para mapear hostname → inbox_id
- Endpoint público POST /track/click para receber eventos de clique
- Leads::AttributionMatcherService para correlacionar clique com conversa
- Integração com IncomingMessageWuzapiService para atribuição automática
- API REST para gerenciar LandingHosts por inbox (index/create/destroy)
- UI: nova aba 'Landing Pages' nas configurações da caixa de entrada
- Dashboard API client dedicado (landingHosts.js)
- RuboCop: refatora shift_signature_name, TrackingController, AttributionMatcherService e WuzapiService
Substitui rails_storage_proxy_url (URL absoluta com host ngrok) por
rails_storage_proxy_path (URL relativa) em file_url e thumb_url.
Problema: ngrok mostra página de interstitial HTML para sub-recursos
carregados pelo browser (img/audio) sem cookie ngrok válido.
O browser recebia HTML em vez da mídia → imagem 'não disponível' e
áudio '00:00/00:00'.
Solução: URL relativa (/rails/active_storage/blobs/proxy/...) resolve
para o servidor atual sem passar pelo ngrok, eliminando o interstitial.
Funciona tanto em localhost:3000 quanto acessando via ngrok no browser.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- MediaHandler: adiciona sanitize_content_type que normaliza audio/opus → audio/ogg
- MediaHandler: detect_extension retorna .ogg (não .mp3) para áudios WhatsApp
- MediaHandler: final_filename força extensão .ogg em áudios que chegam com .mp3
- Attachment: normalize_opus_blob_content_type! agora verifica apenas content_type
(remove checagem de extensão de filename que impedia normalização de blobs .mp3)
- Attachment: audio_metadata chama normalize_opus_blob_content_type! para corrigir
blobs existentes na primeira vez que são acessados (lazy fix)
WhatsApp envia áudio como container OGG/Opus (bytes OggS = 4f 67 67 53),
mas declarava mimetype audio/opus. Browsers não conseguem reproduzir container
OGG via MIME audio/opus — precisam de audio/ogg.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
## Description
Fixes a critical bug where conversations assigned to a team could be
auto-assigned to agents outside that team when all team members were at
capacity.
## Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
## Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> **Medium Risk**
> Changes core assignment selection for both legacy and v2 flows;
misconfiguration of `allow_auto_assign` or team membership could cause
conversations to remain unassigned.
>
> **Overview**
> Prevents auto-assignment from crossing team boundaries by filtering
eligible agents to the conversation’s `team` members (and requiring
`team.allow_auto_assign`) in both the legacy `AutoAssignmentHandler`
path and the v2 `AutoAssignment::AssignmentService` (including the
Enterprise override).
>
> Adds test coverage to ensure team-scoped conversations only assign to
team members, and are skipped when team auto-assign is disabled or no
team members are available; also updates the conversations controller
spec setup to include team membership.
>
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
67ed2bda0cd8ffd56c7e0253b86369dead2e6155. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
Adds an account-level setting `keep_pending_on_bot_failure` to control
whether conversations should move from pending to open when agent bot
webhooks fail.
Some users experience occasional message drops and don't want
conversations to automatically reopen due to transient bot failures.
This setting gives accounts control over that behavior. This is a
temporary setting which will be removed in future once a proper fix for
it is done, so it is not added in the UI.
Upgrade rails to 7.2.2 so that we can proceed with the rails 8 upgrade
afterwards
# Changelog
- `.circleci/config.yml` — align CI DB setup with GitHub Actions
(`db:create` + `db:schema:load`) to avoid trigger-dependent prep steps.
- `.rubocop.yml` — add `rubocop-rspec_rails` and disable new cops that
don't match existing spec style.
- `AGENTS.md` — document that specs should run without `.env` (rename
temporarily when present).
- `Gemfile` — upgrade to Rails 7.2, switch Azure storage gem, pin
`commonmarker`, bump `sidekiq-cron`, add `rubocop-rspec_rails`, and
relax some gem pins.
- `Gemfile.lock` — dependency lockfile updates from the Rails 7.2 and
gem changes.
- `app/controllers/api/v1/accounts/integrations/linear_controller.rb` —
stringify params before passing to the Linear service to keep key types
stable.
- `app/controllers/super_admin/instance_statuses_controller.rb` — use
`MigrationContext` API for migration status in Rails 7.2.
- `app/models/installation_config.rb` — add commentary on YAML
serialization and future JSONB migration (no behavior change).
- `app/models/integrations/hook.rb` — ensure hook type is set on create
only and guard against missing app.
- `app/models/user.rb` — update enum syntax for Rails 7.2 deprecation,
serialize OTP backup codes with JSON, and use Ruby `alias`.
- `app/services/crm/leadsquared/setup_service.rb` — stringify hook
settings keys before merge to keep JSON shape consistent.
- `app/services/macros/execution_service.rb` — remove macro-specific
assignee activity workaround; rely on standard assignment handlers.
- `config/application.rb` — load Rails 7.2 defaults.
- `config/storage.yml` — update Azure Active Storage service name to
`AzureBlob`.
- `db/migrate/20230515051424_update_article_image_keys.rb` — use
credentials `secret_key_base` with fallback to legacy secrets.
- `docker/Dockerfile` — add `yaml-dev` and `pkgconf` packages for native
extensions (Ruby 3.4 / psych).
- `lib/seeders/reports/message_creator.rb` — add parentheses for clarity
in range calculation.
- `package.json` — pin Vite version and bump `vite-plugin-ruby`.
- `pnpm-lock.yaml` — lockfile changes from JS dependency updates.
- `spec/builders/v2/report_builder_spec.rb` — disable transactional
fixtures; truncate tables per example via Rails `truncate_tables` so
after_commit callbacks run with clean isolation; keep builder spec
metadata minimal.
- `spec/builders/v2/reports/label_summary_builder_spec.rb` — disable
transactional fixtures + truncate tables via Rails `truncate_tables`;
revert to real `resolved!`/`open!`/`resolved!` flow for multiple
resolution events; align date range to `Time.zone` to avoid offset gaps;
keep builder spec metadata minimal.
- `spec/controllers/api/v1/accounts/macros_controller_spec.rb` — assert
`assignee_id` instead of activity message to avoid transaction-timing
flakes.
- `spec/services/telegram/incoming_message_service_spec.rb` — reference
the contact tied to the created conversation instead of
`Contact.all.first` to avoid order-dependent failures when other specs
leave data behind.
-
`spec/mailers/administrator_notifications/shared/smtp_config_shared.rb`
— use `with_modified_env` instead of stubbing mailer internals.
- `spec/services/account/sign_up_email_validation_service_spec.rb` —
compare error `class.name` for parallel/reload-safe assertions.
When businesses use WhatsApp Business App (co-existence mode) or
Instagram App or TikTok alongside Chatwoot, messages sent from the
native apps were not synced properly back to Chatwoot. This left agents
with an incomplete conversation history and no visibility into responses
sent outside the dashboard. Additionally, if these echo messages did
arrive, they appeared as "Sent by: Bot" in the UI since they had no
sender, making it confusing for agents.
This PR subscribes to WhatsApp `smb_message_echoes` webhook events and
routes them through the existing service with an `outgoing_echo` flag,
mirroring how Instagram already handles echoes. On the Instagram side,
echo messages now also carry the `external_echo` content attribute and
`delivered` status.
On the frontend, messages with `externalEcho` are distinguished from bot
messages showing a "Native app" avatar and an advisory note encouraging
agents to reply from Chatwoot to maintain the service window.
<img width="1518" height="524" alt="CleanShot 2026-01-29 at 13 37 57@2x"
src="https://github.com/user-attachments/assets/5aa0b552-6382-441f-96aa-9a62ca716e4a"
/>
Fixes
https://linear.app/chatwoot/issue/CW-4204/display-messages-not-sent-from-chatwoot-in-case-of-outgoing-echo
Fixes
https://linear.app/chatwoot/issue/PLA-33/incoming-from-me-messages-from-whatsapp-business-app-are-not-falling
* chore: update scheduled messages author association to nullable and adjust related specs
* chore: update sender handling for WhatsApp messages and add external sender name
* feat: Adds model for scheduling messages
* feat: Implement scheduled message handling and processing jobs
* feat: Add ScheduledMessagesController and associated specs for managing scheduled messages
* refactor: Simplify scheduled message job specs and improve metadata handling
* feat: Add ScheduledMessagePolicy for managing access to scheduled messages
* feat: Add routes for managing scheduled messages
* feat: Add scheduled message event handling and broadcasting
* feat: Add JSON views for scheduled messages creation, destruction, updating, and indexing
* feat: Update scheduled message status and dispatch update event after message creation
* feat: Ensure scheduled message updates trigger dispatch event
* feat: Add mutation types for managing scheduled messages
* feat: Add additionalAttributes prop to Message component and provider
* feat: Implement scheduled message handling in ActionCable and Vuex store
* feat: Add unit tests for scheduled messages actions and mutations
* feat: implement scheduled messages functionality
- Added support for scheduling messages in the conversation dashboard.
- Introduced new components: ScheduledMessageModal and ScheduledMessages for managing scheduled messages.
- Enhanced ReplyBottomPanel to include scheduling options.
- Updated Base.vue to handle scheduled message styling.
- Integrated Vuex store module for managing scheduled messages state.
- Added necessary translations for scheduled messages in English and Portuguese.
* feat: add pagination to scheduled messages index and update tests accordingly
* chore: update scheduled messages specs for future time validation and response status
* chore: enhance scheduled messages API with pagination and add skeleton loader component
* feat: add create_scheduled_message action to automation rule attributes
* feat: implement create_scheduled_message action and enhance attachment handling
* feat: add scheduled message functionality with UI components and localization
* test: enhance scheduledMessages mutations tests with meta handling and structure
* chore: update label to display file name upon successful upload in AutomationFileInput component
* feat: add initialAttachment prop to ScheduledMessageModal and update ReplyBox to pass attachment
* chore: prepend_mod_with to ScheduledMessagesController for better module handling
* fix: attachment visibility in ScheduledMessageItem component
* chore: enhance ScheduledMessage model with validations and reduce controller load
* refactor: simplify ScheduledMessagesAPI methods by removing unnecessary instance variable
* chore: update event emission for scheduled message creation in ReplyBox and ScheduledMessageModal
* refactor: update status configuration to use label keys
* chore: update date formatting in ScheduledMessageItem component
* refactor: collapse logic to checkOverflow and update related functionality
* chore: add author indication for current user in scheduled messages
* chore: enhance scheduled message metadata with author information and localization
* fix: send message shortcut
* chore: handle errors in scheduled message submission
* chore: update scheduled message modal to use combined date and time input
* chore: refactor scheduled messages handling to remove pagination and update related tests
* fix: ensure scheduled messages update status and dispatch on failure
* fix: update scheduled message due date logic and simplify sending checks
* refactor: rename build_message method for send_message
* fix: update scheduled message creation time and improve test reliability
* chore: ignore unnecessary check
* chore: add scheduled message metadata handling in message builder, add scheduled message factorie and update specs
* refactor: use scheduled message factorie creation in specs
* chore: streamline error handling in scheduled message job and remove dispatch logic
* fix: change scheduled_messages association to destroy dependent records
* refactor: remove unused attributes from scheduled message payload builder
* chore: update scheduled message retrieval to use conversation association
* chore: correct cron format for scheduled messages job
* chore: remove migration for author_type in scheduled_messages
* feat: enhance scheduled messages management with delete confirmation and error handling
* chore: set cron poll interval to 10 seconds for improved scheduling precision
* feat: include additional_attributes in message JSON response
* feat: enhance scheduled message validation and localization support
* chore: update scheduled message display
* Merge branch 'main' into Cayo-Oliveira/CU-86aenh268/Mensagens-agendadas
* feat: add scheduled message indicators and validation for message length
* fix: remove unnecessary condition from line-clamp class binding
* feat: update scheduled messages localization and enhance content validation
* feat: update scheduled messages order, enhance scheduledAt computation, and add message association
* fix: reorder condition for Facebook channel message length computation
* fix: change detection for attachments in scheduled messages
* fix: remove unnecessary colon from close-on-backdrop-click prop in ScheduledMessageModal
* chore: add error handling for scheduled message deletion and update localization for delete failure
* fix: enforce minimum delay of 1 minute for scheduled messages and update validation
* fix: remove unused private property and improve locale formatting for scheduled messages
* fix: adjust positioning of DropdownBody in ReplyBottomPanel and clean up schema foreign keys
* docs: add scheduled messages management APIs and payload definitions
---------
Co-authored-by: gabrieljablonski <contact@gabrieljablonski.com>
### What
Forces `account_id` to be applied consistently in queries and message creation paths.
### Why
Some queries were missing `account_id`, leading to cross-account scans and slow performance in large datasets.
### Changes
* Added `account_id` to the relevant query columns.
* Ensured messages are always created within the correct account scope.
* Updated `created_at` handling where required for consistency.
### Impact
* Prevents cross-account queries.
* Improves query performance.
* Reduces risk of incorrect data access across accounts.
### Notes
No functional behavior change for end users. This is a performance and safety fix.
We are expanding Chatwoot’s automation capabilities by
introducing **Conversation Workflows**, a dedicated section in settings
where teams can configure rules that govern how conversations are closed
and what information agents must fill before resolving. This feature
helps teams enforce data consistency, collect structured resolution
information, and ensure downstream reporting is accurate.
Instead of having auto‑resolution buried inside Account Settings, we
introduced a new sidebar item:
- Auto‑resolve conversations (existing behaviour)
- Required attributes on resolution (new)
This groups all conversation‑closing logic into a single place.
#### Required Attributes on Resolve
Admins can now pick which custom conversation attributes must be filled
before an agent can resolve a conversation.
**How it works**
- Admin selects one or more attributes from the list of existing
conversation level custom attributes.
- These selected attributes become mandatory during resolution.
- List all the attributes configured via Required Attributes (Text,
Number, Link, Date, List, Checkbox)
- When an agent clicks Resolve Conversation:
If attributes already have values → the conversation resolves normally.
If attributes are missing → a modal appears prompting the agent to fill
them.
<img width="1554" height="1282" alt="CleanShot 2025-12-10 at 11 42
23@2x"
src="https://github.com/user-attachments/assets/4cd5d6e1-abe8-4999-accd-d4a08913b373"
/>
#### Custom Attributes Integration
On the Custom Attributes page, we will surfaced indicators showing how
each attribute is being used.
Each attribute will show badges such as:
- Resolution → used in the required‑on‑resolve workflow
- Pre‑chat form → already existing
<img width="2390" height="1822" alt="CleanShot 2025-12-10 at 11 43
42@2x"
src="https://github.com/user-attachments/assets/b92a6eb7-7f6c-40e6-bf23-6a5310f2d9c5"
/>
#### Admin Flow
- Navigate to Settings → Conversation Workflows.
- Under Required attributes on resolve, click Add Required Attribute.
- Pick from the dropdown list of conversation attributes.
- Save changes.
Agents will now be prompted automatically whenever they resolve.
<img width="2434" height="872" alt="CleanShot 2025-12-10 at 11 44 42@2x"
src="https://github.com/user-attachments/assets/632fc0e5-767c-4a1c-8cf4-ffe3d058d319"
/>
#### NOTES
- The Required Attributes on Resolve modal should only appear when
values are missing.
- Required attributes must block the resolution action until satisfied.
- Bulk‑resolve actions should follow the same rules — any conversation
missing attributes cannot be bulk‑resolved, rest will be resolved, show
a notification that the resolution cannot be done.
- API resolution does not respect the attributes.
---------
Co-authored-by: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: iamsivin <iamsivin@gmail.com>
Co-authored-by: Pranav <pranav@chatwoot.com>
High-traffic accounts generate excessive database writes due to agents
frequently switching between conversations. The update_last_seen
endpoint was being called every time an agent loaded a conversation,
resulting in unnecessary updates to agent_last_seen_at and
assignee_last_seen_at even when there were no new messages to mark as
read.
#### Solution
Implemented throttling for the update_last_seen endpoint:
**Unread messages present:**
- Updates immediately without throttling to maintain accurate
read/unread state
- Uses assignee_unread_messages for assignees, unread_messages for other
agents
**No unread messages:**
- Throttles updates to once per hour per conversation
- Checks if agent_last_seen_at is older than 1 hour before updating
- For assignees, checks both agent_last_seen_at AND
assignee_last_seen_at - updates if either timestamp is old
- Skips DB write if all relevant timestamps were updated within the last
hour
- Consolidated two separate update_column calls into a single
update_columns call to reduce DB queries
Fixes https://github.com/chatwoot/chatwoot/issues/13097
### Problem
The PR #12176 removed the `before_save :setup_webhooks` callback to fix
a race condition where Meta's webhook verification request arrived
before the channel was saved to the database. This change broke manual
WhatsApp Cloud channel setup. While embedded signup explicitly calls
`channel.setup_webhooks` in `EmbeddedSignupService`, manual setup had no
equivalent call - meaning the `subscribed_apps` endpoint was never
invoked and Meta never sent webhook events to Chatwoot.
### Solution
Added an `after_commit` callback that triggers webhook setup for manual
WhatsApp Cloud channels
## Linear Ticket
https://linear.app/chatwoot/issue/CW-4569/nomethoderror-undefined-method-blocked-for-nil-nomethoderror
## Description
Fixes NoMethodError in ConversationMuteHelpers that occurs during
contact deletion race condition.
When a contact is deleted, there's a brief window (~50-150ms) where
contact_id becomes nil but conversations still exist. If ResolutionJob
runs during this window, the muted? method crashes trying to call
blocked? on nil.Fixes # (issue)
## Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
## How Has This Been Tested?
- Created orphaned conversations (contact_id = nil)
- Called muted?, mute!, unmute! - all return gracefully
- Verified async deletion still works correctly
## Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
Co-authored-by: Sojan Jose <sojan@pepalo.com>
CSAT scores are helpful, but on their own they rarely tell the full
story. A drop in rating can come from delayed timelines, unclear
expectations, or simple misunderstandings, even when the issue itself
was handled correctly.
Review Notes for CSAT let admins/report manager roles add internal-only
context next to each CSAT response. This makes it easier to interpret
scores properly and focus on patterns and root causes, not just numbers.
<img width="2170" height="1680" alt="image"
src="https://github.com/user-attachments/assets/56df7fab-d0a7-4a94-95b9-e4c459ad33d5"
/>
### Why this matters
* Capture the real context behind individual CSAT ratings
* Clarify whether a low score points to a genuine service issue or a
process gap
* Spot recurring themes across conversations and teams
* Make CSAT reviews more useful for leadership reviews and
retrospectives
### How Review Notes work
**View CSAT responses**
Open the CSAT report to see overall metrics, rating distribution, and
individual responses.
**Add a Review Note**
For any CSAT entry, managers can add a Review Note directly below the
customer’s feedback.
**Document internal insights**
Use Review Notes to capture things like:
* Why a score was lower or higher than expected
* Patterns you are seeing across similar cases
* Observations around communication, timelines, or customer expectations
Review Notes are visible only to administrators and people with report
access only. We may expand visibility to agents in the future based on
feedback. However, customers never see them.
Each note clearly shows who added it and when, making it easy to review
context and changes over time.
https://one.newrelic.com/alerts/issue?account=3437125&duration=259200000&state=d088e9b7-d0ce-3fcf-fda5-145df8b9cb2a
## Description
Pass serialized data instead of ActiveRecord object in
dispatch_destroy_event to prevent ActiveJob::DeserializationError when
the notification is already deleted.
This error occurs frequently because RemoveDuplicateNotificationJob
deletes notifications, and by the time the async EventDispatcherJob
runs, the record no longer exists.
## Type of change
- [ ] Bug fix (non-breaking change which fixes an issue)
## How Has This Been Tested?
## Checklist:
- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> Avoids ActiveJob deserialization failures by sending serialized data
for notification deletion and updating the listener accordingly.
>
> - `Notification#dispatch_destroy_event` now dispatches
`NOTIFICATION_DELETED` with serialized `notification_data` (`id`,
`user_id`, `account_id`) instead of the AR object
> - `ActionCableListener#notification_deleted` reads
`notification_data`, finds `User`/`Account`, computes
`unread_count`/`count` via `NotificationFinder`, and broadcasts using
the user’s pubsub token
> - Specs updated to pass `notification_data` and assert payload
(including `unread_count`/`count`)
>
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
e2ffbe765b148fdfd2cd2e031c657c36e423c1f5. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
---------
Co-authored-by: Vishnu Narayanan <iamwishnu@gmail.com>
