> ## Documentation Index > Fetch the complete documentation index at: https://docs.getvector.me/llms.txt > Use this file to discover all available pages before exploring further. # Read check results > Catch up on what your competitors changed, look into a single check, spot the checks that failed, and see what monitoring costs. Once [competitors](/competitors/add-competitors) are under [monitoring](/monitoring/set-up-monitoring), Vector builds up a history of checks and what they found. You read it in **Monitoring**, in the **Track** group of the left menu. ## Catch up on what changed The **Summary** tab opens first. Across the top: * **On monitoring** and **Watched sources** — how many competitors and sources are watched right now. * **Checks · 30d** and **Changes · 30d** — how many checks ran and how many changes they found in the last 30 days. Below, **Latest changes** lists the most recent findings. Click **See all →** for the **Recent changes** tab: everything found in the last 30 days, newest first, one row per change naming the competitor and source, with a link to the proof. **last swept** shows when the last round of checks ran. If nothing has been found yet, the tab says so: **All quiet — no changes in the last 30 days**. ## See what monitoring costs The **Automatic cost · 30d** tile on **Summary** shows what the schedule itself cost in the last 30 days: the AI calls and page fetches of checks nobody started by hand. Under it, a line names where most of that went. **See the full breakdown** opens [AI usage](/ai/ai-usage). If the figure is higher than you'd like, check the competitors you watch least closely less often. ## Look into a check The **Check log** tab is the full history of monitoring, newest first. Each row is a **run**: one round of checks of one competitor's sources, with how many changed. Click **details** on a run to see it source by source: what was checked, whether it came back **changed**, **no change**, **skipped** or **could not read**, and a short summary of what Vector saw. A change also shows how confident Vector is in it. A change held for your review has **Confirm** and **Dismiss** buttons right there. See [Finding confidence and feedback](/monitoring/finding-confidence). ## Spot checks that failed A source that couldn't be read is marked **could not read**. When the runs you're looking at include any, a warning above the log says how many source checks couldn't be read, so a problem is visible without opening every row. A run is marked by its overall result: * **changed** — at least one source changed, even if others failed * **partial error** — some sources couldn't be read, and none of the rest changed * **could not read** — every source failed, so a broken check never reads as a quiet day * **no change** — everything was read, and nothing changed ## Narrow down the log The filters above the log narrow it by competitor, by outcome — **Changed**, **No change** or **Error** — and by kind of source. Set the period with the **7d**, **30d** and **90d** presets or pick dates. Once you've changed a filter, **Reset** goes back to the default: everything from the last 90 days. The calendar above the table has one cell per day, shaded from **less** to **more** by how many checks ran, with days that found a change marked. Click a day to see only its checks; click it again, or **clear day**, to go back. The calendar follows the filters you've set. The log loads the most recent runs first. **Load more** at the bottom brings in older ones. ## Where to go next Turning monitoring on in the first place. Letting Vector act on a change without you.