Your Toolio data arrives through a chain of work: files land, metrics refresh against them, and planning jobs run on the refreshed numbers. The Pipeline Health dashboard puts one day of that work on a single timeline, so you can see what ran, when it finished, and what is still going.

You can also tell Toolio when your data is supposed to be ready. Set a readiness target — a daily deadline for the pipeline, a phase, or one operation — and the page reports whether each day met it.

Open the profile menu in the top-right corner of Toolio and choose Pipeline Health. The page opens on today in your organization's time zone.

Every timestamp on the page — bars, tooltips, deadlines — is shown in that time zone, and the timeline states which one it is used above the chart.

[Screenshot: The profile menu with Pipeline Health highlighted]

Use the arrows on either side of the date button to step back and forward one day, or click the date itself to pick a day from the calendar. Today returns to the current day.

The calendar stops at today. The one exception is an overnight readiness target whose window is already open — when an evening window has opened for tomorrow's deadline, you can step forward one day to watch it.

Use the refresh button in the toolbar to pull in new runs. Deadlines and countdowns keep updating on their own, but the runs themselves are fetched when the page loads and when you refresh.

The timeline is grouped into the three things the pipeline does, in the order they happen:

Each group header shows how many runs it holds. Below it, one row per operation — the thing that ran, such as a specific importer or a specific refresh job.

[Screenshot: The timeline with the three phase groups expanded]

Each bar is one run, drawn from its start time to its finish. Where several runs overlap too closely to draw separately, they are folded into one bar — darker bars hold more runs, and a count appears beside them.

Hover a bar to see the run's status, start and finish times, duration, records processed, error count, who triggered it, and the filename or error reason where there is one. A run triggered by a person shows their email under Triggered by; a run started by the scheduler shows nothing there.

When you are viewing today, a vertical line marks the current time and everything to the right of it is shaded as future. An unfinished run's bar stops at that line rather than running to the edge of the chart.

The five statuses in the legend and the status filter are:

Click the status filter in the toolbar to show only the statuses you care about; each one carries its count for the day. The rows, bars, and counts always describe the same selection, so a filtered timeline is a complete picture of what you selected. Show all statuses clears it.

A refresh operation that covers more than one resource can be expanded — click the arrow beside the operation name to see one lane per resource, each with its own bars. Rows containing a failure or a warning expand on their own when a day first loads, so problems are visible without hunting for them.

Expand breakdowns and Collapse breakdowns in the timeline header open and close every row at once.

Longest run reports the duration of the slowest run in the row, which is what tells you whether a row is slow or simply busy.

vs SLA appears once you have readiness targets configured. It shows how each row did against its deadline, and each group header shows how many of its operations were on time.

A row with runs that have no start time yet shows N awaiting start beside its name. Those runs cannot be plotted, so they are counted rather than drawn.

Click Manage targets to open Data readiness targets. A target has three parts:

Two presets cover the common shapes: Overnight batch (20:00 → 07:00) and Morning refresh (01:00 → 09:00). A new target starts at 01:00 → 09:00. When Window opens is later on the clock than Ready by, the window starts the previous evening, which is what a nightly batch needs.

[Screenshot: The Data readiness targets dialog with a target being added]

Without a window, yesterday's finished refresh would look like today's. Window opens draws the line: a run that finished at or before that time is excluded from the day's assessment, and a run that started earlier but finished inside the window still counts.

It also protects a good result from a later update. If your refresh was ready at 05:00 against a 09:00 deadline, an intraday top-up that starts at 13:00 does not undo that. A first delivery at 10:00, though, is late.

The question-mark icon in the dialog header opens How readiness targets work, which walks through eight worked examples on a shared 01:00 → 09:00 window.

Targets can overlap, and the most specific one applies. An operation target overrides a target on its phase, and a phase target overrides the pipeline-wide one. Each target card notes what it overrides, so you can see the hierarchy at a glance.

This means you can set one pipeline-wide deadline as your default and then carve out the operations that genuinely need a different one.

Use the edit and delete icons on each card to change or remove a target. Removing the pipeline-wide target leaves your phase and operation targets in place.

Once you have at least one target, a Data readiness band appears above the timeline. Its headline counts the operations that were ready on time out of the operations a target measured.

Beside it, counters break down what is left. Missed shows on every day. On today you also get At risk (a deadline within the next 30 minutes and nothing ready yet), In progress, and Not due yet. Click a counter to dim every timeline row it does not describe; click it again to clear the highlight. Finish time unavailable appears only when a run finished without recording when — that one is a readout, not a filter.

When you are looking at today, a Next deadline tile names the next moment something is due and which target owns it.

[Screenshot: The Data readiness band showing counters and the next deadline]

Readiness details opens a panel listing every target, worst result first. Each entry gives the verdict, the deadline, and — for a target covering several operations — how many were on time.

Where something went wrong, the entry names what to look at: Held up by the operation that bound the result, Waiting on one still running, or Slowest lane when a single operation's own lanes are the issue. Click an entry to jump to that row in the timeline.

Verdicts read as plain sentences, and the wording tells you which case you are in:

At the bottom of the panel, Toolio names any target that governed nothing that day. This is not a miss. It happens for two reasons: the scope ran nothing, or a more specific target already judged all of its work.

A target measures whether the work that ran finished in time. It does not require work to run, so a weekly job is not marked late on the six days it was never going to run.

No. The timeline works on its own. Targets add the Data readiness band, the vs SLA column, and the deadline markers on the chart.

No. Targets are measurement only. Job scheduling is configured elsewhere and is untouched by anything on this page.

The scope list is built from recent run history plus your enabled refresh metric models. An operation that has not run recently enough to appear in that history cannot be offered a target yet.

Once a day is over, a deadline that passed with nothing eligible is a result rather than a countdown. On today the same target reads as not due yet until its deadline arrives.

A failure is answered by a successful retry in the same lane. If the retry finished inside the window, that lane delivered and the target is met.

Deadlines and countdowns are recalculated against the live clock, so an in-progress result can become at risk or missed while the page is open. The runs themselves only change when you refresh.

No. Pipeline Health is a day-level view across all three phases at once. For a single run's full detail — row-level import errors, for example — the Data Imports page is still the place to go.