dashboard/availability的端点健康检查是否能设置检查频率或禁用

[lowlandsheperd]

如果不涉及到同一供应商多个节点的切换，是否可以配置较低的“端点健康”频率，或者关闭检查功能？

目前查看日志，即使无活跃的请求，也会每分钟会检查6-8次。

[github-actions]

Thank you for your question.

Based on my analysis of the codebase, the endpoint health check (Endpoint Probe Scheduler) is always enabled and currently does not have an ENABLE switch to disable it. However, you can adjust the probe frequency through environment variables.

Configuration Options

The following environment variables can be used to adjust the probe behavior (see .env.example lines 124-131):

Variable	Default	Description
ENDPOINT_PROBE_INTERVAL_MS	10000 (10s)	Interval between probe cycles
ENDPOINT_PROBE_TIMEOUT_MS	5000 (5s)	Timeout for each probe request
ENDPOINT_PROBE_CONCURRENCY	10	Number of concurrent probes
ENDPOINT_PROBE_CYCLE_JITTER_MS	1000	Random jitter added to each cycle
ENDPOINT_PROBE_LOCK_TTL_MS	30000	Leader lock TTL for distributed coordination

To reduce probe frequency, you can increase ENDPOINT_PROBE_INTERVAL_MS. For example:

ENDPOINT_PROBE_INTERVAL_MS=60000  # Probe every 60 seconds instead of 10 seconds

Relevant Code

• src/lib/provider-endpoints/probe-scheduler.ts (lines 18-31) - Configuration parsing
• src/instrumentation.ts (lines 275-284) - Scheduler startup (always enabled in production/development)

Current Limitation

As noted in .env.example line 126:

Provider Endpoint Probing (always enabled) - There is no ENABLE switch; it is enabled by default.

If you need the ability to completely disable the endpoint probe scheduler, this would require a feature enhancement to add an ENABLE_ENDPOINT_PROBING environment variable.

---

Automated response from Claude AI

[ding113]

您好，参考机器人回答调整 env 即可。