Passive healthcheck support is similar to nginx's fail_timeout

[vislee]

Like me reported issue #77 .
When only passive healthchecks are used, unhealthy nodes should have a time duration for failure.

[CLAassistant]

All committers have signed the CLA.

[Tieske]

Since we have passive and active healthchecks, I think a user can already configure an active healthcheck instead of using a fail_timeout. Set the unhealthy interval in checks.active.unhealthy.interval.

@vislee This covers the usecase no?

[vislee]

Since we have passive and active healthchecks, I think a user can already configure an active healthcheck instead of using a fail_timeout. Set the unhealthy interval in checks.active.unhealthy.interval.

@vislee This covers the usecase no?

@Tieske Thank you for your answers.
This configuration checks.active.unhealthy.interval can bring healthy nodes online in time.
But, In ​cloud API gateway, some test domain name nodes are down for a long time.

[Tieske]

@vislee sorry I didn't get your remark;

But, In ​cloud API gateway, some test domain name nodes are down for a long time.

Can you elaborate on this? what is the issue here?

[vislee]

@vislee sorry I didn't get your remark;

But, In ​cloud API gateway, some test domain name nodes are down for a long time.

Can you elaborate on this? what is the issue here?

@Tieske Sorry, I didn't make it clear.
In ​cloud API gateway, lots of test domain nodes are down for a long time.
This unhealth nodes will cause the active healthcheck to send a request every checks.active.unhealthy.interval to check the node to be healthy (I read the code is this, hope I understand is right), These requests are redundant in actual scenarios.
So, this test domain, passive healthcheck like nginx are enough.