Signals
Spot meaningful movements in your engineering metrics. Signals highlight notable changes across PRs, issues, and team health, so you can see what needs attention, analyze it with AI, discuss it, and act.
Signals is available to all Swarmia users on Signals and on your Swarmia homepage. The feature is currently in beta and actively being expanded.
Overview
Signals answer one question: Has something meaningfully changed?
A signal opens when a metric crosses an absolute or trend threshold, and stays open while it persists. When the condition clears, it closes.
What signals track
Signals cover the metrics that matter most to engineering teams, at both leadership and team level. The supported set is limited today and will keep expanding in the coming months.
Effort/Investment
FTE changes: Sudden staffing shifts (org or team level)
Category trends: Investment categories trending up or down
Delivery
PR cycle time: Slow merge-to-close cycles or outlier teams
PR merge time: Time from open to merge taking longer than usual
PR review rate: Reviews slowing down or backlogs building up
Batch size: Reviews getting backed up with oversized changes
Time to first review: First review taking too long to arrive
Unlinked PRs: Merges without linked issues
Issue cycle time: Issues taking longer than normal to resolve
Team signal trends: Notable changes in metrics across child teams within the organization

Using Signals in Swarmia
For each signal, you can:
View details – See what triggered it, which PRs, issues or teams are involved, plus a link to a view to dig into the data
Analyze with Swarmia AI – Get a natural-language explanation of what changed and why, streaming in real time, then chat to dig into root causes and suggested actions, or save it as an annotation.
Annotate – Add a note or discuss the data with your team.
Resolve – Mark it handled until it fires again. You can unresolve it if needed.
Snooze – Mute the signal for a set time if you already know about it and don't need reminders
AI-Powered Insights
The analysis opens with the signal already in context: what it detected and why it matters, the recent trend, likely root causes and solutions, and surrounding context (for example, "This spike started after the infrastructure push"). Use it to get up to speed on an unfamiliar metric, correlate a signal with what your team was doing at the time, or find less obvious explanations, such as review latency that points to process rather than staffing.
Viewing History & Trends
Each signal shows how the trend changed over time. Look for:
Steady escalation – The problem got worse over time and likely needs intervention
Spikes – Ups and downs; this can signal outliers or competing deadlines
Slow decay – A good sign that the fix is working
To learn from a past change, use the signal details and AI analysis to see what shifted, what likely caused it, and what action helped.
Weekly digest
Stay on top of your metrics with a weekly digest of your signals. Configure them for yourself or your team to receive a weekly notification in Slack or MS Teams channels.

FAQ
"I don't see any signals"
Signals are available to all Swarmia users, but if nothing is active, there may be nothing to show yet.
If you're experiencing issues accessing Signals, reach out to hello@swarmia.com for support.
"I see a signal. Now what?"
Open the signal to see the details: which PRs are stale, what changed in the metric, and when the signal opened
Click "Analyze with Swarmia AI" for a summary
Take action (for example, split large changes or dig into root causes)
Snooze or resolve the signal to keep your view clean
"This signal is a false alarm, or won't go away"
Signals refresh daily, not in real time. If the data has changed, the signal will likely clear tomorrow.
Snooze it to fade it from your view for the time you set (typically 1 week), or Resolve it to clean up your view while the metric catches up. If the condition returns, it will reopen.
Leave a note explaining why (for example, "Team is on sprint planning week"), or click Analyze with Swarmia AI to confirm why it is still open.
"I keep getting the same signal"
If it has been active more than half of the past 90 days, it's likely structural—not a one-off spike.
Look for a process, tooling, or staffing change that would address the root cause
Last updated
Was this helpful?