Notification Channels

Mockarty supports sending notifications through external messaging services — Telegram, Slack, Discord, Microsoft Teams, and others. This allows users to receive real-time alerts about test results, fuzzing findings, contract drift, and system events directly in their preferred messenger.

Start with the setup steps below. If you only want to receive notifications,
you need a channel that an administrator has already made available.

Who sets up notifications

1. An administrator adds a channel in Admin Panel → Notification Channels
   and sends a test message. This is where the service credentials are set.
2. A user opens Settings → My Notification Channels, connects to an
   available channel, and chooses which events to receive.
3. For a shared team channel, a namespace administrator creates a subscription
   under Settings → Namespace Callbacks → Team channels. The subscription
   chooses the events and whether to send them immediately or as a digest.

Mockarty sends each selected event through the connected service. You can check
whether a message was delivered in the delivery log. The sections below show
which services are available and how to connect them.

Closing a subscription form does not cancel a request already sent. Its result is shown while you remain on the same page and namespace. Changing page or namespace closes the form; a late response cannot update the new view. If the result is uncertain, check the team subscription list before creating another subscription.

If the event or channel catalogue cannot be loaded within 15 seconds, the form shows an error and Retry. Retry is manual; opening or closing the form does not repeat a subscription creation request.

Choosing a delivery mode

Choose Immediate to send each team-channel event as it happens. Choose a
digest every 5 or 15 minutes, hourly, or daily to combine events into one
message with counts by event type. Set this when creating the subscription in
Settings → Namespace Callbacks → Team channels (digestMode in the API and
MCP tool). A digest may arrive up to one minute after its window ends.

The delivery log shows digesting while events are being collected. After
sending, it shows delivered or failed with a reason.

Digest messages use the same event titles as the in-app bell. They are in
English by default. An administrator can set MOCKARTY_DIGEST_LANGUAGE=ru for
Russian text. If MOCKARTY_PUBLIC_URL is set, the message links to the full
notifications in /ui/notifications.

If the service is temporarily unavailable, Mockarty retries delivery. Muting
the event, disconnecting the channel, or changing its recipient stops pending
delivery to the old recipient. The delivery log still shows the stopped
attempt.

Supported Channels

Admin Setup

Adding a Channel

1. Go to Admin Panel → Notification Channels
2. Click Add Channel
3. Select the channel type
4. Fill in the configuration (follow the setup hints shown for each type)
5. (Optional) Set Namespace — leave empty to make the channel global (visible in every namespace) or enter a specific namespace name to scope it. The field is safe to change later via update.
6. Click Save
7. Use Test button to verify connectivity

Updating a Channel (secrets-safe)

When editing an existing channel, you don’t need to re-enter tokens, webhook URLs, or other sensitive fields — the UI sends an empty string for fields you leave blank and the server keeps the previously stored value for them. Only fields you actively fill in get overwritten. This lets you rename a channel, flip its enabled state, or move it between namespaces without ever exposing its secrets again.

Admin audit log

Every admin-initiated channel change is recorded in the audit log (visible in Admin Panel → Audit Log). Three action codes identify the events:

Secret fields (bot tokens, webhook URLs, passwords) are never written to the audit log — only the public metadata is captured. Use this log to trace who configured a channel, when it was disabled, or which admin removed a target that stopped receiving alerts.

Channel Configuration Examples

Telegram

1. Open Telegram, find @BotFather
2. Send /newbot, follow the prompts to create a bot
3. Copy the bot token (format: 123456789:ABCdef...)
4. In Mockarty Admin → Channels → Add → Telegram:
   • Bot Token: paste the token
   • Bot Username: @YourBotName (optional, for display)
5. Save and enable

Users will connect by clicking a deeplink that opens the bot in Telegram.

Slack

Option A: Bot Token (recommended for DMs)

1. Go to api.slack.com/apps → Create New App
2. Add Bot Token Scopes: chat:write, users:read.email
3. Install to workspace, copy the Bot Token (xoxb-...)
4. In Mockarty: paste as Bot Token

Users enter their Slack email to receive DMs.

Option B: Incoming Webhook (channel-level)

1. In Slack App → Incoming Webhooks → Add New Webhook
2. Select a channel, copy the URL
3. In Mockarty: paste as Webhook URL

Discord

1. In Discord: Server Settings → Integrations → Webhooks → New Webhook
2. Select a channel, copy the webhook URL
3. In Mockarty: paste as Webhook URL

No per-user binding needed — messages go to the channel.

Microsoft Teams

1. In Teams channel: click ... → Workflows
2. Select “Post to a channel when a webhook request is received”
3. Follow the setup wizard, copy the webhook URL
4. In Mockarty: paste as Webhook URL

Note: Legacy Office 365 Connector URLs are being retired. Use Workflows (Power Automate) webhooks.

PagerDuty

1. In PagerDuty: Services → select service → Integrations tab
2. Add Integration → Events API v2
3. Copy the Integration Key
4. In Mockarty: paste as Integration Key (routing_key)
5. Optional: set Severity — one of critical, error, warning, info (defaults are applied per event type if omitted)

OpsGenie

1. In OpsGenie: Teams → select team → Integrations → Add integration → API
2. Copy the API Key
3. In Mockarty: paste as API Key
4. Select Region — us (api.opsgenie.com) or eu (api.eu.opsgenie.com). Using the wrong region causes silent delivery failures.

Testing a Channel

After saving, click the Test button on any channel to send a test notification. For channels that require a recipient (Telegram, Slack), you’ll need to provide a chat ID or email.

User Setup

Connecting to Channels

1. Go to Settings → My Notification Channels
2. Available channels configured by admin are listed under Available Channels
3. Click Connect on the desired channel

Telegram Connection

1. Click Connect on the Telegram channel
2. A deeplink opens – click it to go to the bot in Telegram
3. Press Start in the bot chat
4. Connection is confirmed automatically

Slack Connection

1. Click Connect on the Slack channel
2. Enter your Slack workspace email
3. Mockarty resolves your Slack user ID and sends DMs directly

Webhook Channels (Discord, Teams, etc.)

These are channel-level — click Connect and you’re done. All users connected to the same channel receive the same messages.

Notification Preferences Matrix

After connecting, configure which events trigger notifications. The preferences matrix lists every channel-deliverable event (dozens of them), grouped by category — test runs, performance, fuzzing, contracts, chat mentions and direct messages, meetings, task tracker, reviews, license and system alerts. Each channel column has an “All” toggle to switch the whole channel on or off in one click, and every event can be enabled or disabled per channel individually.

A few examples of canonical event IDs:

The full, always-current list comes from GET /api/v1/channels/preferences (the eventTypes field) — the UI renders the matrix from it, so new event types appear automatically.

Escalating unread notifications

If a bell notification stays unread for too long, Mockarty can escalate it to one of your connected messengers as a digest. Enable it in Settings → My Notification Channels → Escalation: pick the target channel and the threshold (how long a notification may stay unread before escalation). The feature is opt-in and off by default; enabling it requires at least one connected channel binding.

API: GET /api/v1/notifications/escalation / PUT /api/v1/notifications/escalation.

Notification policy: quiet hours, digests and routing

Settings → Notifications policy sets rules for the whole namespace; a test plan’s Notifications button sets rules for that plan only, and each person can keep their own. When rules conflict the most specific wins (plan, then person, then namespace, then admin); extra channels add up across the levels.

• Muted events and Mute until — the event reaches nobody at this level: no bell, no e-mail, no messenger.
• Quiet hours (UTC) — inside the window your e-mail and messenger notifications are held and arrive as one digest when the window ends. The in-app bell still shows every event at once. With Critical events still fire a critical event comes through immediately.
• Digest — outside quiet hours, e-mail and messenger notifications of the chosen severities are collected and sent as one message every cadence minutes. Critical events are never held. With Minimum events to send the digest waits for that many events — but not longer than a day.
• Severity routing and Always-on sinks — the event is also sent to the chosen channels (for example every critical event to an on-call Slack channel), whatever the recipient’s own quiet hours. A channel receives each event once, however many people it concerned.
• Per-event overrides (set through the API or by an agent with the notification_policy tool) — change an event’s severity, give it its own channels (replace keeps only those), or suppress repeats: with suppress for 10m the same event about the same plan or run reaches a person at most once in 10 minutes.

The digest arrives as the event Notification digest — you can mute or route it like any other event.

Messenger Bots — two-way commands

Telegram, Slack, VK Teams, Mattermost and Rocket.Chat channels are not just outgoing pipes: when the channel is configured for interaction, the bot also accepts commands — the same set on every messenger.

Connecting your account

A bot only acts on behalf of users it knows:

• Telegram — connect from Settings → My Notification Channels (deeplink → press Start).
• VK Teams, Mattermost, Rocket.Chat — click Connect on the channel in Settings → My Notification Channels. Mockarty shows the exact command with a one-time token (valid for a few minutes) and a Copy command button; send it as described below, and the bot answers when your account is linked.
• VK Teams — send the bot /start <token>.
• Slack — request a one-time code via POST /api/v1/channels/slack/connect (or the Connect button in Settings), then send connect <code> to the bot in a DM.
• Mattermost — run /mockarty link <token> in Mattermost.
• Rocket.Chat — send the bot mockarty link <token> (the trigger word, then the command) — in the room the outgoing webhook is attached to, or in a direct message with the bot.

An unknown user who writes to a bot gets onboarding instructions instead of a refusal.

Setting up the Mattermost bot

Mattermost drives the bot over a slash command (no bot needs to stay connected):

1. In the Mockarty admin, add a Mattermost channel and fill Server URL + Slash Command Token (you’ll get the token from Mattermost in the next step).
2. In Mattermost, Integrations → Slash Commands → Add a command with trigger word mockarty, Request URL https://<your-mockarty>/api/v1/messenger/webhook/mattermost/<channel-id>, method POST. Mattermost shows the command Token — paste it into the Mockarty channel’s Slash Command Token and save.
3. Users type /mockarty <command> (e.g. /mockarty ask how do I mock gRPC). Command output is private to the sender; a meeting link is posted to the channel.

The Webhook URL field stays optional — set it if you also want outgoing notifications; the interactive bot works from Server URL + Slash Command Token alone.

Setting up the Rocket.Chat bot

Rocket.Chat drives the bot over an outgoing webhook (Rocket.Chat has no Mattermost-style slash-command integration):

1. In the Mockarty admin, add a Rocket.Chat channel and fill Server URL + Outgoing Webhook Token (you’ll get the token from Rocket.Chat in the next step).
2. In Rocket.Chat, Admin → Integrations → New → Outgoing with event Message Sent, trigger word mockarty, and URL https://<your-mockarty>/api/v1/messenger/rocketchat/outgoing/<channel-id>. Rocket.Chat shows a Token — paste it into the Mockarty channel’s Outgoing Webhook Token and save.
3. Users type mockarty <command> (e.g. mockarty ask how do I mock gRPC). Because outgoing-webhook replies are public, command output posts into the room.
4. Optionally add Bot Auth Token + Bot User ID (a Rocket.Chat personal-access-token for a bot user) to enable the membership guard and call-lifecycle relay via the REST API.

The Webhook URL field stays optional — set it if you also want outgoing notifications; the interactive bot works from Server URL + Outgoing Webhook Token alone.

Commands

In group chats, address the bot as /command@BotName (Telegram) or by mentioning it (Slack @bot, VK Teams). Slack additionally supports slash commands via a configurable endpoint and always verifies request signatures with the channel’s Signing Secret. On Mattermost every command is a subcommand of the slash trigger — /mockarty ask …, /mockarty meet, /mockarty tasks, /mockarty link <token> — and each request is verified against the command’s token.

Voice messages (Telegram, VK Teams)

Send the bot a voice note (or an audio file / round video note) and it transcribes it with the Mockarty speech stack, then acts on the words:

• In a private chat the transcript is treated as a task for the Mockarty AI agent — a hands-free /do. The bot replies with what it heard, hands the task to the agent under your account, and posts the agent’s answer back into the chat when it finishes. When speech synthesis is available on the deployment, the answer is also sent as a voice message (Telegram and VK Teams) — ask by voice, hear the reply. Requires your Telegram to be connected to a Mockarty account.
• In a group chat the bot posts the transcript for everyone. If the group is linked to a discussion (/link), the transcript also joins the buffered history, so /summarize and /topics cover what was said out loud.

Voice notes up to 10 minutes are accepted. Transcription must be available on the deployment (a speech runner online or an external transcription API configured) — otherwise the bot says so honestly.

Linking a group chat to a discussion

A Telegram, VK Teams or Slack group can be linked to a Mockarty discussion: mint a token from the discussion panel (Sync → ), then send /link <token> in the group. From then on the bot keeps the recent history, /summarize posts a digest into the linked discussion, and /topics segments the buffered history into knowledge-base topics.

On Slack, linking additionally requires the app to subscribe to the message.channels event and be invited to the channel — that is how the bot sees the chatter it buffers for /summarize and /topics.

On Mattermost, the chat↔discussion link is /mockarty synclink <token> (distinct from /mockarty link, which binds your account), and buffering needs an outgoing webhook on the channel pointed at /api/v1/messenger/mattermost/outgoing/<Mockarty-channel-id> (use the same token as the slash command). Then /mockarty summarize and /mockarty topics work over the buffered history.

Group moderation

In groups the Telegram and VK Teams bots can guard membership with the full set /guard off | warn | grace | kick. Slack supports /guard off | warn | kick (a bound user sets it; the bot warns — or removes with the channels:manage scope + the member_joined_channel event — an unauthorized joiner); grace is available on Telegram and VK Teams. Mattermost supports /mockarty guard off | warn | kick too, but — because Mattermost gives the bot no join events — it enforces by a periodic membership poll and therefore needs an opt-in bot token (a bot account with manage_channel_members) set in the channel’s Bot Token field; without it guard stays off. Only a chat administrator can change the policy — except on Slack, where any member whose account is connected to Mockarty can, because Slack has no per-channel admin concept the bot can check. In grace mode a newcomer who is not connected to Mockarty gets a welcome message and a time window to connect before being removed; warn only warns, kick removes immediately, off disables guarding. Chat admins are always exempt — the guard never warns or removes an administrator (or the bot itself), so enabling kick will not touch the people who run the chat.

Telegram delivery mode: pull or webhook

By default the Telegram bot polls for updates (works without any inbound connectivity). For lower latency you can switch a channel to webhook mode — Mockarty registers a webhook with Telegram and receives updates on a public endpoint protected by an auto-generated secret:

POST /api/v1/admin/channels/:id/telegram/webhook
{"enabled": true}            # switch to webhook (rotateSecret: true to re-issue the secret)

The mode is also editable in the channel’s settings form (Mode: pull / webhook). Switching back and forth requires no restart.

AI Agent Integration

The AI agent (notifier sub-agent) can send notifications through system channels without credentials:

1. list_system_channels — returns available channels with IDs
2. send_via_channel — sends a notification through a channel by ID

Example agent interaction:

User: "Send a summary of today's test results to Slack"
Agent: [calls list_system_channels → finds Slack channel]
Agent: [calls send_via_channel with channel_id, title, text]

The agent can also use direct tools (send_slack, send_telegram, etc.) with explicit credentials if system channels are not configured.

Without AI Agent

The notification system works entirely without the AI agent:

• Automatic notifications: Events (test completions, fuzzing findings, etc.) trigger notifications through the EventBus → Channel Dispatcher pipeline
• User preferences: Control which events go where via the preferences matrix
• Admin broadcasts: Admin can test channels directly from the admin panel

API Endpoints

Admin API

Delivery reliability

Every channel dispatch is logged to channel_delivery_log with a status that progresses through pending → delivered, or pending → retrying → dead_letter after exhausting max_retries. The admin UI under Notification Channels → Delivery Log gives you two views:

• Dead Letter – deliveries that have used up all retries. You can requeue any entry with one click to send it back through the retry worker.
• Retry Queue – deliveries currently waiting for their next retry attempt, sorted by next_retry_at.

A leader-only background worker picks up due retries, applies exponential backoff, and moves the row to dead_letter once retry_count >= max_retries.

Per-event message templates

The dispatcher renders every notification through an override-aware template engine. Lookup order is (most specific → least specific):

1. (eventType, channelType, lang) – fully-specific override
2. (eventType, channelType, "") – any language
3. (eventType, "", lang) – any channel
4. (eventType, "", "") – any channel, any language
5. Built-in default shipped with the code

Templates use Go text/template syntax and have access to the raw event payload ({{ .Payload.fieldName }}), the resolved title ({{ .Title }}), body ({{ .Body }}), and deep link ({{ .Link }}).

Manage overrides from Notification Channels → Message Templates in the admin UI, or directly via the /api/v1/admin/channel-templates endpoints.

User API

See Also

• Webhooks & Callbacks – Mock-level webhooks for HTTP, Kafka, and RabbitMQ callbacks
• Integrations Guide – Integration tokens, resolver nodes, runner agents, and CI/CD
• Administration Guide – User management, namespaces, and system configuration
• API Reference – Full REST API documentation