You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: content/en/docs/measuring/best-practices.md
+9-1Lines changed: 9 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -152,4 +152,12 @@ We recommend setting `system_check_treshold` to **2** in your production setup o
152
152
### 17. Idle Duration
153
153
If you are trying to calculate an energy per container you should set the `idle-duration` configuration value high enough so you get a stable value to base the offset on.
154
154
155
-
We recommend at least *120 s*.
155
+
We recommend at least *120 s*.
156
+
157
+
### 18. Internal Networking Only
158
+
External networking introduces variable latency in your benchmarks and thus should be avoided whereever possible.
159
+
160
+
One method to achieve this is to set the network to internal only. See [Docker Compose directive](https://docs.docker.com/reference/compose-file/networks/#internal).
161
+
GMT supports this feature natively.
162
+
163
+
In case you need access to external networking we recommend you at least try to cache the request in a warmup run and then run the final benchmark on the cached / internally mirrored result.
0 commit comments