Receiving error "No host available" multiple times an hour when running behind ALB

InfluxDB 1
1

We are running 2 instances of InfluxDB 1.x community edition behind an AWS application load balancer. The load balancer shows all targets to be healthy at all times, yet several times an hour the client in one of our services throw an error “No host available”. The client is used in an api service that processes about 7 requests per second so there is some load on the servers.

I have looked through all available metrics on the influxdb instances but could not find a correlation with the error timestamps. I looked at the number of currently active connections but that did not correlate well with the error appearing.

Is there any metric I can look at to investigate this issue and get to the bottom of it?

2

Hello @jelstner,
You’re wondering about InfluxDB metrics to look at? You can use the free cloud tier to monitor you infuxdb instances.

3

@jelster - Do you know if the requests are making it to the ALB? From what you’re describing of the clients, it almost sounds like they aren’t able to successfully make the requests to the ALB.

I’m pretty rusty on ALB logs, but you might be able to add a tracer header to the requests on your client to see if they are making it to the ALB, along the lines of Request tracing for your Application Load Balancer - Elastic Load Balancing