Skip to content

Commit

Permalink
[doc][ybm] Tablet peer alert (#23942)
Browse files Browse the repository at this point in the history
* Tablet peer alert

* edit

* Update docs/content/preview/yugabyte-cloud/cloud-monitor/cloud-alerts.md

Co-authored-by: Daniel Isen <[email protected]>

* new Azure regions

---------

Co-authored-by: Daniel Isen <[email protected]>
  • Loading branch information
ddhodge and daniel-yb authored Sep 17, 2024
1 parent e4f5943 commit 70aa7d7
Show file tree
Hide file tree
Showing 3 changed files with 17 additions and 3 deletions.
4 changes: 2 additions & 2 deletions docs/content/preview/architecture/yb-tserver.md
Original file line number Diff line number Diff line change
Expand Up @@ -19,11 +19,11 @@ The YugabyteDB Tablet Server (YB-TServer) service is responsible for the input-o

The following diagram depicts a basic four-node YugabyteDB universe, with one table that has 4 tablets and a replication factor of 3:

![Tserver overview](/images/architecture/tserver_overview.png)
![TServer overview](/images/architecture/tserver_overview.png)

The tablet-peers corresponding to each tablet hosted on different YB-TServers form a Raft group and replicate data between each other. The system shown in the preceding diagram includes sixteen independent Raft groups. For more information, see [Replication layer](../docdb-replication/).

Within each YB-TServer, cross-tablet intelligence is employed to maximize resource efficiency. There are multiple ways the YB-TServer coordinates operations across the tablets it hosts.
In each YB-TServer, cross-tablet intelligence is employed to maximize resource efficiency. There are multiple ways the YB-TServer coordinates operations across the tablets it hosts.

## Server-global block cache

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -289,9 +289,11 @@ The following **Azure regions** are available:
- Seoul (koreacentral)
- Ireland (northeurope)
- Norway (norwayeast)
- Doha (qatarcentral)
- Johannesburg (southafricanorth)
- Texas (southcentralus)
- Singapore (southeastasia)
- Stockholm (swedencentral)
- Zurich (switzerlandnorth)
- Dubai (uaenorth)
- London (uksouth)
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ YugabyteDB monitors the health of your clusters based on [cluster alert](#cluste
| Status | Alert | Level |
| :----- | :---- | :---- |
| Healthy | No alerts<br/>[Disk throughput](#fix-throughput-alerts)<br/>[Disk IOPS](#fix-iops-alerts)<br/>[Fewer than 34% of nodes down](#fix-nodes-reporting-as-down-alerts) | <br/>Warning<br/>Warning<br/>Info |
| Needs Attention | [Node free storage](#fix-storage-alerts)<br/>[More than 34% of nodes down](#fix-nodes-reporting-as-down-alerts)<br/>[Memory Utilization](#fix-memory-alerts)<br/>[YSQL Connections](#fix-ysql-connection-alerts)<br/>[CPU Utilization](#fix-cpu-alerts) | Warning or Severe<br/>Warning or Severe<br/>Warning or Severe<br/>Warning<br/>Warning or Severe<br/>Warning or Severe<br/>Warning or Severe
| Needs Attention | [Tablet peers](#fix-storage-alerts)<br/>[Node free storage](#fix-storage-alerts)<br/>[More than 34% of nodes down](#fix-nodes-reporting-as-down-alerts)<br/>[Memory Utilization](#fix-memory-alerts)<br/>[YSQL Connections](#fix-ysql-connection-alerts)<br/>[CPU Utilization](#fix-cpu-alerts) | Warning or Severe<br/>Warning or Severe<br/>Warning<br/>Warning or Severe<br/>Warning or Severe<br/>Warning or Severe |
| Unhealthy | [More than 66% of nodes down](#fix-nodes-reporting-as-down-alerts)<br/>[CMK unavailable](#fix-cmk-unavailable-alerts) | Severe<br/>Warning |

To see the alert conditions that caused the current health condition, click the cluster health icon.
Expand All @@ -74,6 +74,7 @@ When you receive a cluster alert, the first step is to review the chart for the

| Alert | Metric |
| :--- | :--- |
| [Tablet Peers](#fix-tablet-peer-alerts) | Tablets |
| [Disk Throughput](#fix-throughput-alerts) | Disk IOPS |
| [Disk IOPS](#fix-iops-alerts) | Disk IOPS |
| [Node Free Storage](#fix-storage-alerts) | Disk Usage metric |
Expand All @@ -93,6 +94,17 @@ If you get frequent cluster alerts on a [Sandbox cluster](../../cloud-basics/cre

{{< /note >}}

#### Fix tablet peer alerts

YugabyteDB Aeon sends a notification when the number of [tablet peers](../../../architecture/docdb-replication/replication/#tablet-peers) in the cluster exceeds the threshold, as follows:

- Number of tablet peers is 85% of the cluster limit (Warning).
- Number of tablet peers is 100% of the cluster limit (Severe).

If the number of tablet peers in the cluster approaches the limit for the cluster, consider scaling the cluster horizontally by adding nodes, or vertically by adding vCPUs.

For information on scaling clusters, refer to [Scale and configure clusters](../../cloud-clusters/configure-clusters/).

#### Fix throughput alerts

YugabyteDB Aeon sends a notification when the disk throughput on any node in the cluster exceeds the threshold, as follows:
Expand Down

0 comments on commit 70aa7d7

Please sign in to comment.