All systems operational
Last checked
In progress: Fetching Enedis metering data (a delivery point’s load curves) and the client authorisation form are unavailable during the transition to pvSpline’s new Enedis contract. They are planned to return on 1 December 2026.
Everything else in pvSpline works as usual. Until then, enter a project’s consumption month by month, fill it from a standard profile or import it from a bill.
Uptime over the past 90 days, one bar per day (UTC). Incident history
pvspline.fr
Product pages, the user manual and the bilan verifier.
Operational
You are reading this page, so the website is answering.
Last 90 days: 100% uptime. Days affected, newest first: 24 October 2026, degraded performance 1 h; 19 October 2026, degraded performance 1 h; 16 October 2026, degraded performance 1 h.
app.pvspline.fr
Projects, simulations and your account, in the browser.
Operational(response time 1061 ms)
Last 90 days, from our declared incidents only. No incident declared.
api.pvspline.fr
Sign-in, project saves and the web app’s simulation requests.
Operational(response time 580 ms)
Last 90 days: 100% uptime. Days affected, newest first: 25 October 2026, maintenance 22 h 40 min; 24 October 2026, maintenance 1 d; 23 October 2026, maintenance 1 d; 22 October 2026, maintenance 1 d; 21 October 2026, maintenance 1 d; 20 October 2026, maintenance 1 d; 19 October 2026, maintenance 1 d; 18 October 2026, maintenance 1 d; 17 October 2026, maintenance 1 d; 16 October 2026, maintenance 1 d, and 3 more.
What runs behind those services, as AWS records it: the containers on ECS, the Lambda functions, the map tile builder on EC2 and 3-D view streaming on GameLift Streams.
ECS
1 of 1 task running
Operational
Last 90 days: 99.05% uptime. Days affected, newest first: 23 October 2026, degraded performance 3 min; 22 October 2026, degraded performance 3 min; 21 October 2026, degraded performance 8 min; 20 October 2026, degraded performance 3 min; 19 October 2026, degraded performance 7 min; 17 October 2026, degraded performance 7 min; 15 October 2026, outage 5 min; 13 October 2026, degraded performance 2 min; 12 October 2026, outage 7 min; 11 October 2026, degraded performance 3 min, and 23 more.
ECS
1 of 1 task running
Operational
Last 90 days: 99.85% uptime. Days affected, newest first: 23 October 2026, degraded performance 2 min; 22 October 2026, degraded performance 1 min; 21 October 2026, degraded performance 2 min; 20 October 2026, degraded performance 1 min; 19 October 2026, degraded performance 3 min; 17 October 2026, degraded performance 2 min; 15 October 2026, degraded performance 1 min; 13 October 2026, degraded performance 3 min; 11 October 2026, degraded performance 3 min; 9 October 2026, degraded performance 1 min, and 17 more.
ECS
1 of 1 task running
Operational
Last 90 days: 100% uptime. Days affected, newest first: 13 October 2026, degraded performance 3 min; 8 October 2026, degraded performance 3 min; 7 October 2026, degraded performance 2 min.
Lambda
181 invocations, 0 errors (0.0%) in 15 min
Operational
Last 90 days: 100% uptime. Days affected, newest first: 5 September 2026, degraded performance 2 h.
Lambda
No invocations in 15 min
Standby
No invocations in the last 15 minutes; the function wakes on its next request.
Last 90 days: 99.95% uptime. Days affected, newest first: 25 October 2026, maintenance 22 h 40 min; 24 October 2026, maintenance 1 d; 23 October 2026, maintenance 1 d; 22 October 2026, maintenance 1 d; 21 October 2026, maintenance 1 d; 20 October 2026, maintenance 1 d; 19 October 2026, maintenance 1 d; 18 October 2026, maintenance 1 d; 17 October 2026, maintenance 1 d; 16 October 2026, maintenance 1 d, and 3 more.
Lambda
1 invocation, 0 errors (0.0%) in 15 min
Operational
Last 90 days: 100% uptime. Days affected, newest first: 24 October 2026, degraded performance 1 h; 19 October 2026, degraded performance 1 h; 16 October 2026, degraded performance 1 h.
EC2
Not verified
AWS did not answer in time.
Last 90 days, from our declared incidents only. No incident declared.
GameLift Streams
Capacity: 0 warm, 0 in use, 3 at most
Standby
No capacity kept warm: a 3-D session starts on demand, after a short wait.
Last 90 days, from our declared incidents only. No incident declared.
No incidents reported today.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
No incidents reported.
In progress: Fetching Enedis metering data (a delivery point’s load curves) and the client authorisation form are unavailable during the transition to pvSpline’s new Enedis contract. They are planned to return on 1 December 2026.
Everything else in pvSpline works as usual. Until then, enter a project’s consumption month by month, fill it from a standard profile or import it from a bill.
No incidents reported.
No incidents reported.
No incidents reported.
Each service is checked from the website’s own server in Paris when this page is requested, and the result is reused for up to a minute. Reload after that to check again. The web app must load its start page and the cloud API must answer its health check.
These checks confirm that a service is reachable. They sign in to nothing, run no simulation and touch no project, so a service can pass them and still fail at a particular task.
A check refused at the network edge reads Not verified rather than Outage: it says nothing either way about whether the service works for you.
The infrastructure rows are read at the same time from AWS’s records of each component: task counts for the containers, the last 15 minutes of invocations, errors and throttles for the functions, the tile builder’s instance state and the stream group’s capacity. They are reused for up to three minutes. A component idle by design reads Standby. A row the page could not read shows Not verified and leaves the overall status alone, and an outage of an optional component (anything but the simulation engine, the queue worker and the API function) counts as Degraded overall.
Each card’s bars are its last 90 days, one per day in UTC, coloured by the worst it saw that day: green when nothing was wrong, amber for degraded performance, orange for a partial outage, red for an outage, blue for maintenance, grey for a day with no record. They come from two sources: the records AWS keeps and the incidents we declare. For the functions, and the website and cloud API they serve, AWS’s records give the share of invocations that failed in each hour (at least 5 failures making 10% of them is degraded, 50% an outage); for the containers, they give the tasks running against those wanted. The web app, the tile builder and 3-D view streaming keep no such record: their bars show our incidents only, and no uptime.
Uptime is the share of that recorded time without an outage. An outage counts in full, a partial outage for 30% of its duration, and degraded performance and maintenance not at all.
Incidents are the problems we declare ourselves, with their updates: they cover what no check can see as well as what one can. While one is open, the rows it names and the overall status show it, and each one stays in the incident history. Times are in UTC.
Something not working that this page doesn’t show? Tell us. We answer in English or French.