Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | memoryGb => memoryGiB | Henning Baldersheim | 2024-06-11 | 3 | -12/+11 |
| | |||||
* | Merge pull request #31446 from vespa-engine/bratseth/container-ideal-memory | Bjørn Meland | 2024-06-05 | 1 | -1/+1 |
|\ | | | | | Increase container ideal memory load to 0.9 | ||||
| * | Increase container ideal memory load to 0.9 | Jon Bratseth | 2024-06-05 | 1 | -1/+1 |
| | | | | | | | | | | | | | | Vespa will now attempt to allocate memory to this level. If we don't allow it, all allocations will be unfulfiled, since adding more resources won't decrease used memory per container. | ||||
* | | Remove printlns | Jon Bratseth | 2024-06-05 | 1 | -2/+0 |
| | | |||||
* | | Test that we don't rescale when container memory is unfulfiled | Jon Bratseth | 2024-06-05 | 2 | -0/+19 |
|/ | |||||
* | Avoid injecting Exclusivity | Håkon Hallingstad | 2024-05-19 | 1 | -3/+6 |
| | |||||
* | Remove flags dependency from config-model | Håkon Hallingstad | 2024-05-18 | 2 | -14/+9 |
| | |||||
* | Move CapacityPolicies to config-privisioning | Jon Bratseth | 2024-05-12 | 1 | -1/+1 |
| | |||||
* | Make CapacityPolicies independent of NodeRepository | Jon Bratseth | 2024-05-12 | 1 | -1/+4 |
| | |||||
* | Lower ideal query cpu | Jon Bratseth | 2024-04-24 | 3 | -59/+59 |
| | | | | 80% has observed to cause problems, so lowering to 75%. | ||||
* | Add host-memory-ratio flag and add flavor dimension to host-memory | Håkon Hallingstad | 2024-04-22 | 1 | -1/+1 |
| | |||||
* | Avoid Optional with CloudAccount | Håkon Hallingstad | 2024-04-18 | 3 | -8/+6 |
| | |||||
* | Propagate cloud account to ResourcesCalculator | Håkon Hallingstad | 2024-04-17 | 3 | -5/+9 |
| | |||||
* | Store multiple resource suggestions | Ola Aunronning | 2024-01-12 | 2 | -4/+4 |
| | |||||
* | Store GPU metrics and load separately | Martin Polden | 2023-12-20 | 5 | -31/+35 |
| | |||||
* | Revert "Expose GPU load in nodes API" | Harald Musum | 2023-12-19 | 5 | -35/+31 |
| | |||||
* | Store GPU metrics and load separately | Martin Polden | 2023-12-19 | 5 | -31/+35 |
| | |||||
* | Remove unused parameters | Martin Polden | 2023-12-08 | 1 | -1/+0 |
| | |||||
* | Let distributor metric decide cluster stability | Martin Polden | 2023-12-06 | 1 | -1/+20 |
| | |||||
* | Revert "Evaluate flags referenced multiple times at the start of prepare" | Håkon Hallingstad | 2023-11-07 | 3 | -10/+6 |
| | |||||
* | Include NodeRepository, ApplcationId, and ClusterSpec as parameters | Håkon Hallingstad | 2023-11-05 | 3 | -10/+7 |
| | |||||
* | Rename to AllocationParams | Håkon Hallingstad | 2023-11-05 | 2 | -5/+5 |
| | |||||
* | Rename to ClusterAllocationParams | Håkon Hallingstad | 2023-11-04 | 3 | -9/+9 |
| | |||||
* | Evaluate flags referenced multiple times at the start of prepare | Håkon Hallingstad | 2023-11-04 | 3 | -4/+11 |
| | |||||
* | Update ↵ | Bjørn Meland | 2023-11-03 | 1 | -1/+1 |
| | | | | | node-repository/src/test/java/com/yahoo/vespa/hosted/provision/autoscale/AutoscalingTest.java Co-authored-by: Jon Bratseth <bratseth@gmail.com> | ||||
* | Clarification of message | bjormel | 2023-11-02 | 1 | -1/+1 |
| | |||||
* | Don't autoscale single node clusters | bjormel | 2023-11-02 | 1 | -3/+1 |
| | |||||
* | Update copyright | Jon Bratseth | 2023-10-09 | 15 | -13/+15 |
| | |||||
* | Propagate GPU resources when determining current resources for autoscaling | Valerij Fredriksen | 2023-09-28 | 1 | -0/+19 |
| | |||||
* | Update AwsNodeTypes with GPU instances | Valerij Fredriksen | 2023-09-28 | 1 | -11/+62 |
| | |||||
* | Compare advertised resources in flavor satisified check in tests | Valerij Fredriksen | 2023-09-28 | 3 | -37/+51 |
| | |||||
* | Use at least 8GB memory nodes in unit tests | Valerij Fredriksen | 2023-09-08 | 1 | -3/+1 |
| | |||||
* | Adjust by ideal | Jon Bratseth | 2023-08-16 | 1 | -19/+19 |
| | |||||
* | Don't use BCPInfo to scale down | Jon Bratseth | 2023-08-16 | 1 | -1/+16 |
| | |||||
* | More accurate accounting of redistribution cost | Jon Bratseth | 2023-08-16 | 1 | -30/+1 |
| | |||||
* | Test account for change cost | Jon Bratseth | 2023-08-15 | 2 | -31/+97 |
| | |||||
* | AllocatableClusterresources -> AllocatableResources | Jon Bratseth | 2023-08-15 | 2 | -9/+5 |
| | |||||
* | No functional changes only: Wrap current resources in model | Jon Bratseth | 2023-08-15 | 2 | -13/+16 |
| | |||||
* | Consider switching cost when choosing resources | Jon Bratseth | 2023-08-15 | 1 | -1/+1 |
| | |||||
* | Leave 15% headroom when scaling down | Jon Bratseth | 2023-07-31 | 2 | -6/+6 |
| | |||||
* | Set aside enough headroom rather than not scaling down at all | Jon Bratseth | 2023-07-25 | 2 | -6/+6 |
| | |||||
* | Set aside space for headroom in one pass | Jon Bratseth | 2023-07-25 | 1 | -3/+3 |
| | |||||
* | Don't scale down if we are likely to scale back up | Jon Bratseth | 2023-07-20 | 3 | -10/+41 |
| | | | | | | | | | When we decide to scale number of nodes down we'll see a lower ideal load because we need to be able to handle one node going down. This may lead us to be closer to ideal (at current peak load) than otherwise anticipated, such that we are quite likely to soon scale back up. This checks for that and avoids scaling down dimensions where this is the case. | ||||
* | Merge pull request #27787 from vespa-engine/bratseth/no-scalingDuration-min | Jon Bratseth | 2023-07-14 | 2 | -29/+49 |
|\ | | | | | Remove scaling duration min for content clusters | ||||
| * | Remove scaling duration min for content clusters | Jon Bratseth | 2023-07-14 | 2 | -29/+49 |
| | | | | | | | | This has been a safety measure, which doesn't seem necessary any more. | ||||
* | | Deduce tight group limits when using group-size | Jon Bratseth | 2023-07-13 | 1 | -1/+1 |
|/ | |||||
* | Don't pick resources which are against recommendations | Jon Bratseth | 2023-07-07 | 1 | -22/+66 |
| | |||||
* | Test warning | Jon Bratseth | 2023-07-05 | 1 | -1/+12 |
| | |||||
* | Reduce minimum content scaling duration from 8 to 6h | Jon Bratseth | 2023-06-13 | 2 | -7/+6 |
| | |||||
* | Reduce min content scaling duration frpm 12 to 8 | Jon Bratseth | 2023-05-24 | 2 | -10/+10 |
| |