Fix: SQLite lock contention causing db read timeout.

[abishekve]

Summary

Fixes an issue where /api/status/live remained in a pending state due to SQLite database lock contention caused by concurrent monitoring result writes.

Root Cause

• Multiple monitoring workers were writing to check_result_raw concurrently.
• SQLite allows only a single writer at a time.
• Large check_result_raw table size increased query latency.
• Read queries such as fetching latest timestamps were blocked by active writes.
• This caused frontend requests to timeout after 30 seconds.

Changes

• Serialized monitoring result writes using a background write queue.

• Reduced concurrent SaveChangesAsync() contention.

• Enabled SQLite WAL mode for improved read/write concurrency.

• Added SQLite busy timeout handling.

• Added indexes for:

  • check_result_raw(Ts DESC)
  • check_result_raw(EndpointId, Ts DESC)

• Ensured scoped DbContext usage per monitoring operation.

Result

• Prevents frequent SQLite Error 5: database is locked
• Improves /api/status/live responsiveness
• Reduces API timeout occurrences
• Improves monitoring write stability under load