Uptime by weekday

• Uptime is 99.50% across the record — 5.2h of it not fully operational.
• The process behind this service restarted 239 times in the last 30 days. A restart completes far inside the agent's 30-second sampling interval, so it leaves no mark on the percentage above — it is counted here instead, because "nothing was ever recorded against it" and "it never went away" are not the same claim.
• 2 of the last 30 days were only partly recorded, the thinnest at 33% of the day. Those percentages are true of the hours the record was watching, not of the whole day, and their columns are hatched to say so.
• 15 status changes in the last 30 days. That is flapping rather than outages — a probe timeout or heartbeat threshold tight enough to trip on ordinary jitter will produce exactly this shape.
• Slowest around 19:00 at 76 ms, against 62 ms at 05:00 — steady across the day.