Skip to content

Web Intelligence

WebI Live shows every Web Intelligence Processing Server (WIPS) in every environment against its configured limits, the calls running on it now, long and stuck calls, and pile-ups as they form.

Image of the WebI Live page with the collector lamps, tiles, Fleet matrix and Fleet over time

../../images/web-console/insights/webi-full.png

What it answers

  • Is any WebI server close to its session, document cache or memory limit?
  • What is running long right now, and on which server?
  • Is a call stuck, compared with how long that step normally takes?
  • Is a pile-up of long calls forming on one server?
  • What was running at the time of a user complaint?
  • Which documents are the slowest, and in which processing step?

The page only reads. Nothing on it changes a BusinessObjects server, setting or running report.

Open it

In the BusinessObjects console, select Reliability > WebI Live. The i button next to the title opens a field guide that explains every section.

Filters

Filter Values Effect
Environments one button per environment Limits every section.
All WIPS / n WIPS selected pick from a list, or click matrix rows (ctrl or shift to add) Narrows Ceiling headroom, Running now, Timeline and the lists below to the selected servers.
Range 15m, 1h (default), 3h, 6h, 12h, 1d, 3d, 7d Time range of the charts and lists.
Auto-refresh off, 10s, 30s Default 10 seconds for ranges up to 6 hours, off for longer ranges.
UTC toggle Shows every time in UTC instead of browser time.
% of ceiling / absolute segmented control on Fleet over time Plots values as a share of each server's limit or as raw values.
How this page decides button Shows the thresholds the page uses.

Panels

Panel What it shows
Collector lamps One lamp per environment under the title. Green: watching normally. Amber: BusinessObjects server statistics are more than 15 minutes old. Red: calls were counted by BusinessObjects that MetaMiner did not see, usually after a WIPS restart; while red, an empty Running now list does not mean nothing is running. Also states whether limits were read live from the CMS.
WebI servers reporting Servers with current statistics, out of all servers.
Nearest ceiling The server closest to any limit (sessions, document cache, memory against Upper threshold).
Calls in flight now Calls running now, with long, stuck and pile-up counts.
Memory Analysis OFF when Enable Memory Analysis is off on any WIPS; memory thresholds are then not enforced.
BO versions Distinct BusinessObjects versions across the servers.
Restarts in range WIPS process restarts (process id changes) in the range.
Fleet matrix One row per WIPS, all environments, sorted by an attention score. Excel and CSV export a snapshot.
Fleet over time One line per WIPS, cursor synced across charts.
Ceiling headroom Gauges per server: Sessions, Document cache, Memory (RSS) with Lower, Upper and Maximum bands, Threads. Shows the six servers with the highest attention score, or the selected ones.
Running now Calls seen in the last 20 seconds: running for, start, category, method or phase, document, user, WIPS, environment and vs normal (ratio to normal duration). Client polling is excluded.
Pile-up radar Groups of 3 or more calls on one WIPS in the same method and phase, each over 1 minute. When there is none, names the last one seen in 7 days.
Timeline Calls by category (Refresh, Open, Render, Edit, Save / export, Swap, Other) and session counts for the selected servers. Marks show restarts, configuration changes made in MetaMiner and threshold counters. Drag across a chart to pick a window.
What was running Every call in flight in the window you dragged, with start, last seen, duration, category, document, user and WIPS.
Slow documents Documents with sampled calls in the range: p50, p95, maximum, top phase, users, last seen, daily p95 and the audited refresh p95.

Image of the Ceiling headroom cards with the memory bands and one Memory Analysis OFF badge

../../images/web-console/insights/webi-ceilings.png

Image of Running now with one call marked stuck and the Pile-up radar beside it

../../images/web-console/insights/webi-running-pileups.png

Image of the Timeline with a dragged window and the What was running list filled

../../images/web-console/insights/webi-timeline.png

Thresholds

Mark Rule
Long Call running 1 minute or more (red from 5 minutes).
Stuck Call running more than 3 times the 95th percentile for that step on that environment over the last 7 days, and at least 1 minute.
Pile-up 3 or more long calls on one WIPS in the same method and phase.
Gauge colour Amber from 80% of the server's own limit, red from 95%.

Calls are sampled every few seconds, so very fast calls are often not seen and p95 values run high.

TODO: document what must be installed or enabled for the WebI call collector and the process memory figures

Source to check: product owner. The page reads sampled WebI calls, BusinessObjects server metrics and process metrics measured on the BO host; the prerequisites for each are not stated in the web app.

Export

Export the Fleet matrix snapshot and the lists with Excel or CSV. See Insights Dashboards.