Skip to content

Monitor

The Monitor web app charts CPU, memory, Java heap and disk of the MetaMiner Server (MMS) and the size of the MetaMiner Database (MMDB), so you can see whether the server keeps up with your tenant.

Before you begin

  • The MMDB is PostgreSQL. Server metrics are sampled into it by default (server property metrics.pg.enabled, default true).
  • The MMS port is reachable only from your administrators' network. Monitor does not ask users to sign in. See Network Exposure.

Turn it on

Monitor is off by default.

MMA label MMA page Server property Default
Pulse (/web/monitor) WebApps, group Web apps web.monitor.enabled false
  1. In the MetaMiner Administrator (MMA), choose Preferences > Settings... and select the WebApps page.
  2. Tick Pulse (/web/monitor) and save. The change takes effect without a restart.

Open it

  • In the MetaMiner Client (MMC) choose File > Launch Pulse..., or
  • open https://<mms-host>:<port>/web/monitor/mms in a browser.

Monitor has two pages, BusinessObjects and MetaMiner Server, switched in the top bar. In a Power BI deployment use MetaMiner Server. The BusinessObjects page charts SAP BusinessObjects hosts and stays empty ("No BusinessObjects host has reported through the MetaMiner agent yet.").

Time range and refresh

Control Values Default
Time range 5m, 15m, 1h, 6h, 24h, 7d, 30d, 180d 15m on the MetaMiner Server page
Refresh Refresh off, Every 5s, Every 10s, Every 30s, Every 1m Every 5s

Ranges up to 6 hours show raw samples (Java heap every 5 seconds, disk every 30 seconds). Longer ranges show hourly rollups. The range and refresh are kept in the page address. Refresh pauses while the browser tab is hidden.

Status badges always carry a word, not only a colour: Normal, Elevated, High or No data.

MetaMiner Server page

Image of the Monitor MetaMiner Server page, range 24h: key figures row, CPU load and JVM heap charts

../images/dashboards/monitor-mms.png

Key figure What it shows Elevated / High at
MMS process CPU CPU of the MMS process, mean and maximum over the range 50 / 80 percent
Host CPU CPU of the whole MMS machine 60 / 85 percent
JVM heap Heap used, committed and maximum 75 / 90 percent
Host memory Physical memory used of installed 85 / 95 percent
Server volume free Free space on the MMS disk, with a projection of days to full (after 6 hours of samples) 80 / 90 percent used
MMDB size Size of the MetaMiner Database, with its growth per day -
Metrics store Space the server metrics take in the MMDB -
Rollup job Time since the metrics rollup job last ran, and its status -
Panel What it shows
CPU load MMS process and whole host, bucket maximum
JVM heap Used against committed; a dashed line marks the maximum heap
Host physical memory Used against installed RAM
Server volume free Free GB, bucket minimum
MMDB size GB over time
Metrics pipeline How far the rollup has progressed, where raw samples start, the newest sample per feed, and the last three rollup runs with their status

What to look for in a Power BI deployment

Pattern Likely cause Next step
JVM heap stays near the maximum, with regular CPU peaks The heap is small for the tenant size Performance and Memory
MMDB size grows steadily by large amounts per day Activity and capacity history accumulating Data Retention
Server volume free falls quickly MMDB growth, or large temporary files Performance and Memory

Troubleshooting

The page says "This web app is turned off on this MetaMiner Server"

Tick Pulse (/web/monitor) on the MMA WebApps page.

Rollup job shows Not run yet

Ranges over 6 hours need the rollup job. Check that the MetaMiner Server has been running long enough for its first run, and read the last message in the Metrics pipeline panel.

All figures show No data

Server metrics sampling is off (metrics.pg.enabled=false) or the MMDB is not PostgreSQL. Remove the setting and restart the MetaMiner Server.