Monitoring Node's Health

Nethermind 1.9.51 version comes with a pre-packed Nethermind.HealthChecks.dll plugin that allows you to monitor your Nethermind node better. It leverages the power of AspNetCore.Diagnostics.HealthChecks. It simply adds an/healthendpoint to the JSON RPC service which can be used to check the Nethermind's liveness - verify if the node is synced and has at least one peer. Useful when you don't want to query the node before it's able to provide you data available only for fully synced nodes like eth_getBalance.
The Nethermind.HealthChecks.dllplugin will be automatically loaded on Nethermind start.
Enabling and configuring Health Checks
The health checks need to be additionally enabled which can be done either through --HealthChecks.* flags or by adding a "HealthChecks" section to the config file. 
HealthChecks config section example
1
  "HealthChecks": {
2
    "Enabled": true,
3
    "WebhooksEnabled": true,
4
    "WebhooksUri": "https://slack.webhook",
5
    "UIEnabled": true,
6
    "PollingInterval": 10,
7
    "Slug": "/api/health"
8
  }
Copied!
JSON RPC Service needs to be enabled in order for health checks to work --JsonRpc.Enabled true
Each configuration option is described here: 
HealthChecks module
Enabling Health Checks without UI 
1
./Nethermind.Runner --HealthChecks.Enabled true
Copied!
The health endpoint is now available at localhost:8545/health  by default (if your --JsonRpc.Port is 8545). The health endpoint can be configured via --HealthChecks.Slug parameter e.g. --HealthChecks.Slug /api/health. We can if it is working with curl:
1
// Request
2
curl localhost:8545/health
3
​
4
// Example of response for Unhealthy node
5
{"status":"Unhealthy","totalDuration":"00:00:00.0015582","entries":{"node-health":{"data":{},"description":"The node has 0 peers connected","duration":"00:00:00.0003881","status":"Unhealthy","tags":[]}}}
6
​
7
// Example of response for Healthy node
8
{"status":"Healthy","totalDuration":"00:00:00.0015582","entries":{"node-health":{"data":{},"description":"The node is now fully synced with a network, number of peers: 99","duration":"00:00:00.0003881","status":"Healthy","tags":[]}}}
Copied!
Unhealthy returns 503 (Service Unavailable) status code
Healthy returns 200 status code
Enabling Health Checks UI
1
./Nethermind.Runner --HealthChecks.Enabled true --HealthChecks.UIEnabled true
Copied!
Enabling UI will expose an additional endpoint /healthchecks-uiand will allow seeing node's health on a nice UI. To view the UI simply go to http://localhost:8545/healthchecks-ui.
Unhealthy status reported on UI page
Enabling Slack reports
We may also add Slack Webhook endpoint to which our node's health will be reported. We need to pass the --HealthChecks.WebhooksEnabled true and add  the --HealthChecks.WebhooksUri which can be found in your Slack app configuration.
1
./Nethermind.Runner --HealthChecks.Enabled true --HealthChecks.UIEnabled true --HealthChecks.WebhooksEnabled true --HealthChecks.WebhooksUri https://hooks.slack.com/
Copied!
If your node will be Unhealthy you should receive a message similar to this:
Unhealthy
with description of why the node is unhealthy, node's name and information about the machine on which the node is running.
When it becomes Healthy (synced and with peers) you should receive:
Healthy
HealthChecks for producing and processing blocks
In version v.1.10.18, we introduced additional behavior for health checks. We added two fields for HealthChecks config: MaxIntervalWithoutProcessedBlock and MaxIntervalWithoutProducedBlock. The node will return unhealthy status if the interval elapsed without processing or producing a block. Let's use the below config as an example. If the node doesn't process a block for 15 seconds, we will return unhealthy status. Analogically, we will be waiting 45 seconds for a newly produced block.
HealthChecks config section example
1
  "HealthChecks": {
2
    "Enabled": true,
3
    "WebhooksEnabled": true,
4
    "UIEnabled": true,
5
    "Slug": "/api/health",
6
    "MaxIntervalWithoutProcessedBlock ": 15,
7
    "MaxIntervalWithoutProducedBlock": 45
8
  }
Copied!
If we don't set those fields in a config application will try to use them based on seal engine specification. You can see concrete values in the tables below. If we have infinity time, we can still report unhealthy status if our processing or producing threads stopped.
Seal engine
Processing interval
Producing interval
Clique
4 * Period
2 * Period * SingersCount
Aura
4 * StepDuration
2 * StepDuration * ValidatorsCount
Ethash
180
Infinity
Custom/NethDev/None
Infinity
Infinity
health_nodeStatus
In version v.1.10.18, we've also introduced health checks via JSON RPC requests. To do that, we should have HealthChecks.Enabled set to true.
Request
Response
1
{ "jsonrpc":"2.0","method":"health_nodeStatus","params":[],"id":67 }
Copied!
1
{
2
    "jsonrpc": "2.0",
3
    "result": {
4
        "healthy": false,
5
        "messages": [
6
	        "Stopped processing blocks",
7
            "Node is not connected to any peers"
8
        ]
9
    },
10
    "id": 67
11
}
Copied!

Ethereum client - Previous

Plugins

Next - MEV

MEV Plugin

Last modified 5mo ago

Copy link

Contents

Enabling and configuring Health Checks