Background: need to know the traffic between Environment ActiveGate and Dynatrace Managed Clusters, the figure will be used to predict the Internet bandwidth requirement after migrating Dynatrace Managed Clusters to Dynatrace SaaS.
Managed Cluster: 5 x server nodes (hkldmp04, hkldmp05, hkldmp06, hkldmp07, hkldmp08)
Environment ActiveGate: 15 x servers across multiple regions
Actions:
- Installed OneAgent on some of the Env AG and monitor the metric “outgoing traffic” --- was told it represents the traffic rate between AG to Managed Cluster (= 5 server nodes)
- Collect network bandwidth information from Zurich network team, only from AU at this moment.
- Collect snapshot of network activity from windows “task manager”, only from JP at this moment.
Findings:
- AU AG <> Cluster nodes traffic over 24 hours as below
- JP AG <> Cluster nodes traffic over 2 sec is as below
- 2 x JP AG – outgoing traffic collected by OneAgent over 24 hours
- From above findings #1 & #2, it seems Env AG will communicate and/or send data to 5 x Cluster nodes instead of sending data to a single node, as the traffic rate to/from all Cluster nodes are similar.
Questions:
- Regarding the data transfer between AG and Cluster nodes in findings #1 and #2 above, are those data was replicated or distributed to 5 x Cluster nodes?
- For findings #3, does that mean the Internet bandwidth requirement between JP and Dynatrace SaaS can be estimated as avg. 4Mb/s x 2?
- This may be affected by the outcome of question #1.
Featured Posts
