The Loghound panel has 17 views, each split into pages you reach from the navigation on the left. Every view behaves the same way: the page appears at once, each card fills in on its own, and a card that fails takes only itself down.
Who came and how long they really stayed; the access log as it is written. Overview, Live.
Sessions, Map (a card of Networks), Signed-in visitors (a card of Engagement), Where they came from, Pages, Site search, Engagement, When they come.
SEO Tools, Bot forensics, Attacks, Fingerprints, Networks, Performance, Websites, Storage & bandwidth.
Eighteen cards, from log sources to start over. Settings, card by card.
A period picker (today so far by default), a website selector and the filters scope every view at once, and all of it lives in the URL, so any view you are looking at can be bookmarked or shared.
- The page never waits for Solr. Headings, captions and empty states are drawn first; every number arrives afterwards in its own request.
- Each card says what it is doing while it loads, shows the seconds elapsed after five seconds, and says it is still working at twenty.
- A failed card shows its own error with a Retry button. Retrying runs that one request again, not the whole page.
- Errors are diagnoses. A timeout, an unknown hostname, a refused connection, rejected credentials and a missing index are told apart, because each has a different fix.
- Every number says what it counts. A card names its population, for example “all scored sessions in the selected range”.
Set as ui.query_timeout in the configuration.
Deliberately longer, so a slow query reports its own diagnosis instead of being cut off with a generic message.
Only the two Loghound indexes on your Opensolr account, plus the Opensolr account itself for things such as your plan’s storage and bandwidth. There is no general way through the panel to Solr: every query is built on the server from fixed parts, and the browser only sends a view name, a period, chosen filter values and search text.
The panel
- How the panel works
- Long tasks run as jobs
- Demo data
- Known limitations
- Filters & facets