# Fleet on GKE (ECK) behind a Google LB 502 errors **URL:** \ **Category:** Elastic Agent\ **Created:** [February 14, 2024, 8:36pm UTC](https://discuss.elastic.co/t/fleet-on-gke-eck-behind-a-google-lb-502-errors/353309 "2024-02-14T20:36:40Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![trudesea](https://avatars.discourse-cdn.com/v4/letter/t/e480ec/32.png) [@trudesea](https://discuss.elastic.co/u/trudesea)\ **Post date:** [February 14, 2024, 8:36pm UTC](https://discuss.elastic.co/t/fleet-on-gke-eck-behind-a-google-lb-502-errors/353309/1 "2024-02-14T20:36:40Z")
`Preformatted text`o we run Elastic on GKE using ECK. We are on version 8.12. I just installed the Fleet server yesterday with 4 replicas and they are sitting behind a GCP classic http LB. The servers can be seen on the fleet page and have not gone offline. CPU/RAM are well below limits. So I added about 5 agents to test out everything and I noticed they go offine randomly and come back for some time and then offline again. From the agent logs I see: ```auto 09:58:03.111 elastic_agent [elastic_agent][warn] Possible transient error during checkin with fleet-server, retrying 10:00:59.077 elastic_agent [elastic_agent][error] Cannot checkin in with fleet-server, retrying ``` There are quite a few of the retrying messages but it eventually connects. I see hundreds of 502 errors on the LB. It is setup exactly the way my APM server, Elastic and Kibana LBs are configured and they have no issue. Any ideas? The error is kind of vague. I did try to set the affinity to client ip but no luck. Thanks ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [March 13, 2024, 8:37pm UTC](https://discuss.elastic.co/t/fleet-on-gke-eck-behind-a-google-lb-502-errors/353309/2 "2024-03-13T20:37:37Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.