Control Seat V0.2.2
V0.2.2 fixes two things you'd have noticed from using Control Seat this week: charts on slowly-changing values (valve states, mode selections, alarm counts) were sometimes drawing empty, and the Tags panel was flickering during its periodic refresh. Both are fixed.
What's in V0.2.2
Charts of "on-change" values now draw correctly
If you plot a tag whose value only changes occasionally — a valve position, a discrete state, an alarm count that stays flat for hours — a chart of the last hour used to look empty when the last change was longer than an hour ago. The chart had no points to show, because none had arrived in the window, even though the value was well-defined the whole time.
Now the chart pulls the last known value from before your window and uses it as the starting point. A valve that opened three hours ago and hasn't moved since now shows as "open" for the full hour, instead of an empty chart.
This applies everywhere charts pull history: Trends, OpenBridge Line/Bar/Area/Radar, and the Custom Chart.
Tags panel: no more refresh flicker
The Tags panel refreshes tag counts and folder state periodically in the background. If you had rows expanded or scrolled to a specific tag when a refresh landed, they'd sometimes vanish for a beat before coming back, because the refresh was replacing the loaded rows with the count-only snapshot.
The refresh now preserves whatever you had loaded. You won't see a flicker, and rows you expanded stay expanded.
Tag creation and management, faster to work with
We refactored the Tags panel internals into smaller pieces to make the create/edit/move/delete flows easier to iterate on. No user-facing change today, but expect a noticeably tighter Tags panel over the next couple of releases — several long-standing pain points (fewer clicks to create a tag, better search, keyboard nav) will show up here first.
System tags now report the right numbers on cloud tenants
If you dropped CPU Usage, Memory Usage, or Uptime Seconds from the built-in System tag folder onto a dashboard, cloud tenants used to see the whole underlying host's numbers — the shared physical server, not your tenant. That made the CPU chart useless (always low), the memory chart useless (always the same across every tenant on the box), and Uptime meaningless (told you how long the AWS instance had been up, not Control Seat).
All three now report the correct per-tenant values:
- CPU Usage — percentage of your tenant's CPU budget, not the underlying host's.
- Memory Usage / Max Memory — your tenant's memory limit and current usage.
- Uptime Seconds — how long Control Seat itself has been running, matching what Ignition's
[System]Gateway/UptimeSecondsreports.
On-prem installs are unchanged — you were already seeing the machine's actual numbers.
There's also a new CS_SYSTEM_NAME environment variable so cloud deployments can set a friendly gateway name ("Plant 4 Gateway") instead of showing a random container hostname. On-prem still defaults to the machine hostname.
System folder pins to the top of the Tags panel
The built-in System diagnostics folder now sorts above your own folders at every level of the tag tree, so you don't have to scroll past a hundred user folders to find it.
Smaller fixes
- Chart axis rendering — a subtle bug in how the Line chart resolved its Y-axis limits on partial data has been fixed. Axis labels no longer clip on very small ranges.
- Historian routing — internal fixes on how the query router hands off to individual store backends. No user-visible change; enables the "on-change" support above cleanly for all three store types (ClickHouse, Postgres, Timescale).
That's V0.2.2. The chart fix is the one to feel — open any dashboard with a discrete or rarely-changing value and it should just look right.