Reliability Snapshot is the page you screenshot for the weekly review. SLO burn rate per service, error budget remaining per service, MTTR by tier, incident count by week, and a one-line recommendation per service. Everything that matters this week, on one screen.
Four tiles at the top: SLOs at risk this week, average error budget remaining, this-week MTTR, this-week incident count. Each tile has a delta vs the prior week so improvement (or regression) is visible without context. Click any tile to drill into the underlying breakdown.
Below the tiles, each service shows a one-line recommendation: tighten this SLO target, add capacity here, run a postmortem on that week's spike, retrain the prediction model for that service. The recommendations are generated from the underlying signals; they are concrete, not generic.
The snapshot freezes every Friday at 5pm local time. The frozen snapshot is what shows up in the Monday review meeting and what gets emailed to leadership. Live numbers are still visible on the page, but the meeting numbers are stable so a Saturday incident does not retroactively change the slide.
Leadership and product partners often want the snapshot but should not see raw signals or run actions from it. The page has a read-only "executive view" with the snapshot, deltas, and recommendations, but no drill-ins to incidents or runbooks. One link, two audiences.
Subscribe to Nova AI Ops on YouTube for demos, tutorials, and feature deep-dives.
Stop assembling slides on Friday. The snapshot is the slide.