Supports collect Nebula Graph cluster metrics through NGQL statements

[LiuTianyou]

The NebulaGraph monitoring currently supported by HertzBeat collects performance indicator data through the node's open monitoring HTTP API, which is usually for a single node and does not include cluster information.We can support collecting more data through NGQL statement query.

Metrics that can be queried include but are not limited to:

Cluster level:

• Charset and Collation used by the cluster
• Number of meta nodes (ONLINE/OFFLINE) in the cluster
• Number of graph nodes (ONLINE/OFFLINE) in the cluster
• Number of storage nodes (ONLINE/OFFLINE) in the cluster
• The number of sessions connected to the cluster
• Number of queries being executed
• Number of background jobs being executed
• Number of graph spaces in the cluster
• Meta cluster leader information
• ...

Graph space level:

• Basic information of graph space
• label, number of nodes
• The number of background jobs running in the current graph space
• The number of indexes in the current graph space
• The number of jobs being executed in the current graph space
• ...

Any data that can be queried using NGQL or NebulaGraph-compatible OpenCypher syntax

Custom monitoring template

Users can use the NGQL protocol in custom templates to connect to the NebulaGraph cluster and query the indicators they care about.

About licenses

Nebula java client uses the Apache 2.0 licenses

[tomsun28]

This looks very good, we can define a new protocol for Nebula Graph to implement it. 👍👍
As you say, the new protocol can query any data from NebulaGraph-compatible system.