Status Overview

The status page provides a quick view of what VDBs are configured, what servers are attached to them, and their current health status, along with what clients are connected to them.

For each VDB, the data sources will be shown, the last HB time an update about that server was received, the current qps, connections and if configured, replica lag is.

For the clients, the pin option allows a given client to be saved and always viewed on the status dashboard. This is to allow the current state of clients that only periodically to be viewed. If the pinned option is not set, then clients not seen for over five minutes will disappear from the status view.

Note: A client MAY show as red if it is using connection pooling, despite traffic being active from the client.

If in the VDB configuration Cluster manager via cache is enabled and cluster cache has been used (i.e. Hazelcast or Redis working in cluster mode), the cluster information table will appear, showing cache type, status, list of nodes, partitions owned by each node and the cache uptime.

Note: The Hazelcast instance on the Heimdall Manager should not be responsible for owning any partitions; rather, all partitions ownership will given to the Heimdall proxy nodes.

VDB Actions

When the collapsed menu button is clicked, there should be displayed additional actions showed in the image below.

Below are some actions and details what each of the button is used for:

  • Restart All Proxies: Restart all proxy instances for VDB, on the management server only.
  • Clear Authentication Cache: Clear all authentication related Caches (users, passwords, AD groups etc.).
  • Clear Query Cache: Clear Query Cache, if using a shared cache, this may clear all objects in the cache.

Proxy Actions

When the collapsed menu button is clicked, there should be displayed additional actions showed in the image below. Below are some actions and details what each of the button is used for:

  • Clear Credentials Properties: It will clear the credentials properties cache for a specific proxy. It's usefully when the SQL driven authentication is enabled.
  • Heap dump: It will make a proxy heap dump, which will be available in Logs tab. You will need jmap to perform this action
  • Stack trace: It will make a proxy stack trace, which will be available in Logs tab. You will need jstack to perform this action
  • Packet Capturing: You can initiate a timed packet capture on specific Proxy nodes. This includes limiting duration, total number of files, and the size limit for the capture. Port 8087 and 8443 will be filtered out of the capture automatically.
  • Terminate Process: It will terminate and restart proxy process - if the proxy is manged by central manager. Otherwise, if you have an independent proxy running, proxy process will be terminated (not restarted).
  • Network Disconnect: It will disable the network interface of proxy simulating its unavailability. To enable proxy again turning the interface on manually and restarting the proxy is required. Superuser privileges required. Does not work on local proxies, as it's the same as central managers.

The status color for the proxy is handled by the double-check mechanism. This means that for the proxy to turn red, the last two checks must fail. This eliminates issues such as temporary network overload and prevents the proxy from flickering if the connection problem between the proxy and the central manager is only temporary. Hovering over the status ball will display this information.

Query Distribution

The Query Distribution part of VDB preview provides a quick view of some data aggregated during the proxy runtime. You need to enable certain options to view some of them.

  • Writer - percentage of all queries that did go to writer.
    This statistic is enabled by default.

  • Reader - percentage of all queries that did go to reader due Read/Write split.
    To enable this statistic Read/Write Split must be enabled and configured properly.

  • Cache - percentage of all queries that hit the cache.
    To enable this statistic Caching must be enabled and configured properly.

  • Connection Reuse - A ratio of how many connections were opened on the front vs on the back.
    To enable this statistic connection pooling must be enabled and configured properly.

These statistics can be reset by restarting the proxy in the Status tab.