A few improvements to Status Pages

We’ve shipped a bunch of new features to make managing your status pages even easier, including:

Using the agent, or @incident, to draft status page updates for you
Add Pingdom uptime metrics to your status page
Self-serve localization for your page in a bunch of new languages

Plus we’ve polished up the admin page to make it more aligned with our new dashboard look.

Agent supports updating status pages

The agent now supports the ability to update your status pages. It drafts status page updates based on the context of the incident, including the title, message and affected components.

Don’t worry, the agent does not have the ability to publish those updates, because a mistake here could have serious consequences. However, this should reduce some of the workload on your responders by just having them simply review and publish.

Note that this only applies to external status pages for now.

Status page metrics

We've shipped support for displaying Pingdom uptime metrics on your status page. In the settings section of a status page, customers can now add metrics panels under the new Metrics sidebar item.

Customers can choose to display either an uptime or latency time series, which will show up on their status page with a time resolution picker of 24h/7d/30d.

Check out help docs here.

Self-serve status page localization

We’ve now added the ability to self-serve localization for status pages under settings (Basic Settings -> Customization). This will apply across the full status page, including any sub-pages.

You can also choose different languages for each sub-page, which is useful if you have sub pages for different regions.

We currently have support for Portuguese (Brazilian), Japanese, Spanish, French, German, Finnish, Swedish, Japanese, Italian, as well as the default of English (US).

Investigations

Feedback on investigations for declined and merged incidents

At the end of an incident we post a message in the channel asking for feedback on an investigation. Until now that only happened when an incident was resolved or closed. It now also happens when an incident is declined or merged.

An investigation can be just as useful (or not) when declining or merging an incident. There is nothing to set up, it works in both Slack and Teams, and it's live now for everyone with Investigations enabled.

Investigations autonomy graph improvements

We’ve improved the autonomy graphs on the Investigations homepage so they are a bit easier to read and trust. We've updated the design to show parent medians for time saved by default so customers can easily check the calculation for themselves.

The breakdown of the subcategories is still available by expanding the sections.

Investigations dashboard filtering

By popular demand, you can now filter the Investigations dashboard based on the incidents they're investigating — this includes the incident categorization, participants and custom fields.

Additionally we've added some quick actions so you can jump to viewing your own, or your teams’ incidents.

Nexus

Telemetry support for Coralogix

Investigations customers can now connect Coralogix as a telemetry source provider! Customers who use Coralogix can connect their account to give us access to their Logs, Metrics, and Traces.

Customers can tell us during setup which datasets/tiers they use for their Logs/Traces and will query the right ones based on what we are looking for.

Check out our help docs here.

Telemetry support for Splunk

Investigations customers can now connect Splunk as a telemetry source provider for Nexus and Investigations! We support two of their products:

Splunk (Enterprise / Cloud Platform): their core platform that gives us the ability to access Logs, Metrics and Dashboards. We support both cloud and on-prem deployments. See docs here.
Splunk Observability Cloud: their APM product that used to be SignalFx that gives us access to Metrics and Dashboards. See docs here.

On-call

Repeat on a level of an escalation path

Escalation paths can now send repeated notifications on a single level, before moving on to the next. For example:

Page the Infrastructure schedule up to 3 times, before trying the next level

Previously this meant having multiple identical levels in an escalation path which wasn’t that elegant and it meant that "escalate to the next level" wouldn't work as expected.

Scroll and zoom on schedules

We've just released an update to schedules so that you can now horizontally scroll to pan the timeline, and pinch (or hold cmd/ctrl while scrolling) to zoom in and out.

This also includes a number of other little polish pieces:

You're no longer limited to the existing time periods of two weeks, one week, etc. And, once you stop scrolling, we update the URL to match your scroll position, meaning you can share links to the exact view you're seeing with other people
This works on the calendar view too, where you can vertically scroll, and we now overlay month changes in the calendar to make it clearer what days belong to what month
The 'holidays' row now dynamically grows and shrinks depending on the number of holidays you have in that time period

Response

Linear projects as follow-ups or incident tickets

You can now attach a Linear project as a follow-up, or use them as incident tickets. This is useful if you want to do a bigger piece of work after an incident, which would naturally be a whole project in Linear.

You can also create a project to represent the incident. If you do, any follow-ups you create or attach that don't already have a project will be automatically linked to that project.

Multiple transcription languages for Scribe

Customers can now tell us if multiple languages are spoken on their calls, which we use as a strong hint to Scribe to help it guess correctly.

Pull Scribe transcripts from our API

You can now download incident Scribe transcripts from our public API, exactly the way we display them in the dashboard, with timestamps and the person speaking. The endpoints are paginated to support the often very long transcripts.

To do this, you first choose an incident and grab the ID of it. Then you hit the /call_sessions endpoint to get a list of call sessions, pick one, and pass that ID to the /call_transcript_entries endpoint to get a paginated list of entries. A little guide for how to use it can be found here in the docs.

Streams for Microsoft Teams

If you're an Microsoft Teams customer, you can now use streams from the dashboard, which means:

You can have groups of people on different Scribe calls progressing an incident forward
Those groups can also track actions, add attachments and share updates via our dashboard (or within the Microsoft Teams 'tab' in the main channel). Updates will still be shared to the main teams channel, just like they do in Slack

Note, we do not support separate channels for streams at the moment.

Revamped Teams Insights dashboard

We’ve given a bit of love to the Teams Insights dashboard, so you can better understand how incident work is spread across your teams:

Where incident effort is being carried. Total time spent in incidents, split by team or by user, so you can see which teams are carrying the load week to week.
How much of that work happens out of hours. Working, late and overnight hours are broken out separately, so you can spot teams shouldering on-call burden.
Whether post-incident work is getting done. Follow-ups and post-incident tasks created over time, plus how long each team takes to close them, so you can see where learnings are stalling.

With this little glow-up you’ll get:

Results sorted by top user who did the most 'work'
Grouping smaller data. Before, users who did small amounts of work would not show up properly because their portion of the data was so small. Now we have grouped these together into an "Other" section so you can search through them and still get access to underlying data!

What else we've shipped

New

On-callOur Workday, Bamboo HR and HiBob integration are now out of beta and are fully available to Enterprise customers
ResponseExport follow-ups to a Notion database
ResponsePII redaction applies to Scribe transcripts
ResponseWorkflows can now react to the Slack message that triggered them with a specified emoji.
ResponseWorkflows can now include an "Accept incident" step, so you can automatically move a triaged incident into an active status once certain criteria are met
ResponseWhen exporting postmortems to Notion, you can choose to exclude custom field columns if you don't find them useful
ResponseYou can now merge a private incident into another private incident from the dashboard, without the previous public-first workaround briefly exposing it more broadly than intended — unblocking proper security incident response workflows.
PlatformYou can now also filter to Leading incident on the macOS app, so you only see incidents where you hold the lead role.
PlatformExpressions are now available in a drawer to give you a bit of extra space

Improvements

On-callPagerDuty schedule and escalation policy imports now resolve the full list of items fresh at import time instead of relying on the once-daily sync, so changes made shortly before an import are no longer missed. This makes large-scale schedule migrations far more reliable, especially for organizations with thousands of schedules.
On-callSchedule version history now shows which external schedule (e.g. PagerDuty or Opsgenie) a schedule was originally imported from.
On-callThe schedule timeline is now a continuous, zoomable canvas instead of a fixed page you flip through a week at a time. Pan with a two-finger swipe, and pinch to zoom (or hold ctrl/cmd + scroll, or click and drag on the top of the timeline)
On-callThe working hours warning on escalation paths no longer fires for paths built entirely on "Everyone on schedule" or "Everyone on rotation" targets, since those modes page by membership and aren't affected by working hours gating.
On-callWorkflows can now reference who resolved an alert, so you can route follow-ups, say thanks, or feed the result into on-call reporting.
On-callEscalation messages now clearly say when another responder acknowledged first, instead of saying that the escalation expired.
ResponseThe "mark post-mortem as complete" post-incident task now shows who completed it, bringing it in line with the other post-incident tasks that already display this information.
ResponseYou can now exclude incidents by reference number (including multiple references at once) when filtering Insights and the incident list, so you can exclude anomalies that skew your metrics.
ResponseUpdated the muting functionality for the agent in a Slack thread so that it does not get undone by a later one-off tagged request, such as escalating the incident. The bot now performs the requested action and goes quiet again, rather than treating it as an invitation to rejoin the conversation.
ResponseAdded a "Mute in this thread" option to the overflow menu on the agent’s Slack messages, making it easier to discover how to stop the bot from replying in a thread.
NexusYou can now use AWS Organizations' cross-account observability (OAM) with the direct AWS telemetry integration, not just the Grafana-proxied CloudWatch path. This makes it easier for customers with linked AWS accounts to onboard telemetry without extra workarounds.
NexusYou can now edit an existing self-hosted GitLab installation to change its network routing (public or through a private connector) or update its access token, without needing to remove and re-add it. Self-hosted GitLab setup in the Nexus code onboarding wizard also now supports private connector routing.
Status PagesYou can now create retrospective (historical) status page incidents through the public API, with a full reconstructed timeline of past updates and component impacts.
PlatformCatalog entry attribute tooltips now show the derived path, making it clear where a derived attribute's value actually comes from.
PlatformThe "Team hierarchy" view can now be expanded to fill the browser window, giving you more room to work with larger team structures.

Bug fixes

On-callUsers granted a team role in a parent team were sometimes unable to use that permission in child teams in the dashboard.
On-callFixed a bug where the tooltip shown when creating a cover request was being clipped behind the timeline
On-callFixed a bug that could block saving changes to a multi-rotation on-call schedule (e.g. changing a rotation's cadence and reordering people), which previously failed with a "Rotation rank must be unique" error when a recommended rollout was applied.
On-callFixed a bug where accepting a shift swap could incorrectly fail with a "shifts have already passed" error even when both shifts were weeks in the future. The error now only appears when a shift has genuinely passed, with a clearer message when a shift is no longer held due to another schedule change.
On-callFixed the example Alertmanager configuration snippet shown when setting up an Alertmanager alert source — it can now be copied and applied directly without triggering a config validation error.
On-callFixed the MCP schedule_show and schedule_list tools returning a next_handover_at value that could disagree with the schedule's real next shift boundary. Added new next_shift_starts_at and current_shift_ends_at fields to the schedule_show response to reliably show the next shift boundaries per rotation.
On-callWe fixed an issue where WhatsApp notification rules weren't shown in the "On-call notifications" icon column on the team Members page. WhatsApp now appears alongside phone, SMS, and push as its own icon, correctly reflecting whether each user has a high-urgency WhatsApp rule configured.
On-callFixed a Terraform provider bug where setting operation = "one_of" on a String alert attribute condition caused every apply to fail, because the API's normalization of the value to contains_one_of wasn't reconciled against the plan. Terraform applies with this operation now converge correctly.
On-callFixed a bug where asking the agent to request on-call cover from a specific user could instead unilaterally assign them via a schedule override, without their consent. The agent now correctly sends a cover request for the person to accept or decline.
On-callLinks inside the expanded email alert preview now open in a new tab instead of being blocked, fixing an issue where clicking a link (e.g. "View Dashboard") in the embedded email body showed "content blocked."
ResponseFixed an issue where Notion post-mortem exports could fail when custom field values produced duplicate option names.
ResponseFixed an issue in the incident Participants list where a revoked participant's name would collapse to a single character next to the "Revoked" badge. Names now truncate properly with a tooltip showing the full name on hover.
ResponseFixed misaligned labels and clipped values for long catalog-based custom fields in the incident Properties panel. Long names and values now truncate consistently with a tooltip showing the full text on hover.
ResponseFixed a bug where clicking Continue in the Create custom field modal could clear the form and silently fail to create the field, with no error message shown to explain why.
ResponseUnlinking a debrief from an incident now also detaches the associated calendar event, preventing duplicate calendar bookings and unexpected "detached" notifications in the incident channel later on.
ResponseFixed an issue where pressing Enter to confirm a Japanese (or other IME) text conversion would submit the form instead — affecting the AI assistant chat, timeline comments, and inline rename fields. Enter now correctly confirms the conversion first.
ResponseFixed a bug where selecting "No value" on a required custom field (with the no-value option enabled) would fail validation when declaring an incident from Slack, blocking incident creation. This now matches the existing behavior on the web dashboard.
ResponseFixed a bug where clearing a timestamp (e.g. "Fixed at") from the web dashboard's Update timestamps modal could be blocked by a "durations will become invalid" warning, even though clearing it is a valid state. Clearing timestamps in the web dashboard now behaves consistently with the Slack /inc timestamps dialog.
ResponseFixed a bug where pasting or quoting text containing "@incident.io" (for example, forwarding a message that mentions the bot) caused the agent to treat it as a direct tag and jump in uninvited. Quoted and blockquoted mentions are now ignored, just like mentions inside code blocks, so the agent only responds when it's genuinely addressed.
InvestigationsFixed a bug where every monitored Slack channel showed a blank "last activity" timestamp in Investigations → Messages, even though the channels were actively indexed. The underlying query that powers this field has been rewritten so activity timestamps now display reliably.
NexusNexus can now correctly query Datadog logs stored in the Flex storage tier. Previously, Flex-only log indexes returned no log results.
NexusFixed an issue where an OpenSearch data source running in compatibility mode could not be connected.
NexusFixed index pattern entry timing out for data sources with frequently-rotating index names in some formats.
NexusFixed an issue where successful Loki metric queries run against older Grafana instances could be incorrectly reported as failed.
NexusFixed a bug causing OpenSearch log queries to fail with syntax errors at a high rate for some customers.
Status PagesFixed a bug on internal status pages where using an expression to filter your components list would reload the page instead of saving your filter conditions.
Status PagesStatus page history now shows the same period navigation controls on mobile as on desktop, so visitors can page back through older incidents from their phone instead of only on desktop.
PlatformFixed a bug where Slack file attachments containing non-UTF-8 bytes failed to save entirely. These attachments are now sanitized and stored correctly instead of being dropped.
PlatformFixed a bug where the agent could attempt to generate alt-text for SVG images it can't actually process, sometimes producing inaccurate, hallucinated descriptions. SVG images are now correctly skipped instead.