Source states
Every source in Picasi has a state that determines its behaviour on the next collection run. The state shows up as a badge in the source overview.
Active
Default state. The source is fetched on the regular collection interval; new updates land in the Inbox.
Checking in background
Transitional state right after adding a new source or a new channel. Picasi checks technical reachability and metadata. Usually takes a few seconds to minutes. The source stays in this state without collecting updates until the check is finished.
Broken
The last collection attempt failed with a hard error — the target URL doesn’t respond persistently (HTTP 404, DNS error, domain gone). Affected channels stop being fetched until the error is fixed.
What you can do: In the channel detail view you can see the last error. Check the URL in a browser. If the target content still exists but the URL changed, add the channel again with the new URL. If the source is gone, delete the channel or the whole source.
Degraded
Collection failed multiple times in a row (temporary errors, timeouts, or rate limits), but the situation is not hopeless. Picasi keeps fetching the source, but at a longer interval — instead of hourly, e.g. every six hours. Once a collection run succeeds again, the state automatically flips back to Active.
What you can do: Nothing. The state is just a hint that the source is unstable. Treat it as an early warning — if it stays for long, the source may really be dormant.
Quarantined
The source has repeatedly delivered very large or problematic content — e.g. thousands of items in seconds (typical after a feed bug) or redirect loops. Picasi stops fetching the source until an admin manually releases it or support clarifies the cause.
What you can do: Open the channel detail view. The reason for the quarantine is shown there. After clarification, support flips the source back to Active.
Reset the error counter
For Degraded and other error-prone states, the source detail view shows an error counter. An admin can reset it manually (Reset error counter in the channel menu). Useful when a known server issue on the target has been fixed and you don’t want to wait for the next regular attempt.
State transitions
| From | To | Condition |
|---|---|---|
| Checking in background | Active | First successful check |
| Checking in background | Broken | Check failed with a hard error |
| Active | Degraded | Multiple collection runs in a row with timeout or 5xx |
| Degraded | Active | Next collection run succeeds |
| Degraded | Broken | Error becomes persistent (e.g. sustained 404) |
| Active | Quarantined | Unusual behaviour detected |
| Broken | Active | Reset error counter, or channel re-added |
| Quarantined | Active | Only via support / admin intervention |
States in MCP responses
Over MCP the state is exposed as a status field on the source objects. Values: active, pending_check, broken, degraded, quarantined. Details: MCP tools.