datomic 2019-11-29 | Slack Archive

in datomic cloud (prod topology) I'm getting errors in the lambda cloudwatch logs. how fatal are these?

{:cognitect.anomalies/category :cognitect.anomalies/unavailable, :cognitect.anomalies/message "Connection reset by peer", :clojio/throwable :.IOException, :clojio/socket-error :receive-header, :clojio/at 1574976823963, :clojio/remote "10.213.37.146", :clojio/queue :queued-handler, :datomic.ion.lambda.handler/retries 0}

onetom08:11:07

how come the latest (`569-8835` Nov 27th) version of datomic cloud solo only supports N.Virginia region?

$ curl -s  | jq .Mappings.RegionMap
{
  "us-east-1": {
    "Datomic": "ami-066145045fc7a4ad0",
    "Bastion": "ami-09e416d6385c15902"
  },
  "us-east-2": {
    "Datomic": "",
    "Bastion": ""
  },
  "us-west-2": {
    "Datomic": "",
    "Bastion": ""
  },
  "eu-west-1": {
    "Datomic": "",
    "Bastion": ""
  },
  "eu-central-1": {
    "Datomic": "",
    "Bastion": ""
  },
  "ap-southeast-1": {
    "Datomic": "",
    "Bastion": ""
  },
  "ap-southeast-2": {
    "Datomic": "",
    "Bastion": ""
  },
  "ap-northeast-1": {
    "Datomic": "",
    "Bastion": ""
  }
}

marshall18:11:00

AWS marketplace issue. Use the latest version listed in the datomic docs releases oage

👍 4

marshall18:11:02

Page

Oleh K.17:11:03

I wasn't able to create today a Solo stack on new AWS account with the latest template (of 27 Nov), it just fails. Is it a known problem?

marshall17:11:54

@okilimnik use the latest version listed on the datomic docs releases page

marshall17:11:15

The newest on AWS has several problems that we are attempting to resolve

Oleh K.17:11:30

@marshall thanks!

bartuka21:11:27

hi, I would like to understand a little better the behavior of datomic under load. For example, what is happening on the index-mem-mb metric in this situation?

Linus Ericsson20:12:02

Datomic seems to keep the most recent transactions in memory to batch write them efficiently. If I understand correctly, it means that the transactions only store the tx-data for each transaction (and can batch even that) which means Datomic eventually has to re-calculated its indexes. This is probably what can be seen in the last part of the graph, where the JVM obviously does a lot of work GCing, and then, when the index memory is full, recalculated the db and empties its batched up index memory.

bartuka21:11:57

I noticed that for sometime, my application was processing 200msg/s from rabbit and after 15 min of operation it went down to 10msg/s I am looking at the metrics to figure out the mechanics here

2019-11-29

Channels