Server 500 error and no power data or energy data on dashboard

Others having similar issues?

Me too

Yes. Only data I can get right now is live from the “home” part of the app

Yes, bug report submitted.

Happening too often, they need to look at their infrastructure

Virgin Media are having trouble this morning. I guess that might be related?

Same issue for me. No data showing for today

No data recorded at all today... this is becoming far too common to be acceptable now.

I just hit refresh and can see some data from 7:35, so the early shift must be in and trying to fix it

Virgin had what looks like dns problems this morning, seems to be coming back up now. Lots of sites other than givenergy affected too.

When people are saying this is becoming far too common to be acceptable.......how many times have you been affected this year? Is it just me being lucky or you being unlucky?

Ra I've seen it 3-4 times in last month as a newbie I'd say.

Working OK since 07:40

So does this mean all the data for the affected period is effectively lost?

Eco friendly Solar powered servers don’t you know?
Back up and running at sunrise…

Looking at the number of ISPs with issues, this might well be a more general www problem, rather than a specific GE server issue.

straycat264 , the 5 minute interval data for the pretty power and energy charts, yes the data is lost, however the cumulative meter data is still counting and shows the today and total values correctly once the internet link comes back.

I have a Virgin cable link at home which was down for a period early this morning.

The internet connection came back and I have Import Today readings showing my car charge of 25.3 kWh overnight for example.

I'm surprised the inverter doesn't log some of the data locally as a buffer and then upload it to the server when it connects. It isn't a hard thing to do.

Months ago the fragility of the GE server infrastructure was discussed on here but if anything it's getting worse not better. It's really not difficult to run robust systems with auto-failover and redundancy these days and outages like this should be near non-existent. This isn't a hobby app GE, thousands depend on it! Sort it please. No doubt the competitors will have another field day too.

This was issued on a server down notice.

There were internal communications issues, which have been investigated and resolved.

Russ I agree a buffer of 24 hours in the inverter would be useful to cover comms and central server issues

TheDragon (GivEnergy) Why are we seeing too frequent issues please Paul? For an arguably "critical" system it seems fragile and folk, me included, are worried. Not everyone has the facility to run third-party controls and they shouldn't need to. Is some type of resilience/hardening project in the works please?

Simon_C same here 7:38

not just me then.. my data is only live from 730ish.. battery seemed to charge ok though?

I don't know if it is a coincidence, but my Octopus account is also having difficulties - the running total screens for the week are all down, but historical data is still available. Could this be a bigger internet issue?

I have data loss too... I also pointed out in a bug report to GivEnregy that the overnight charge of the battery did not take place - battery is currently 35% whereas it should be nearer the 80% requested charge.
I have asked them if the charging can actually happen if there is a Givenergy server outage? I would hope that these sort of things were controlled localy.
Anyone else with similar omission of overnight battery charge?

Simon_C , the 5 minute interval data for the pretty power and energy charts, yes the data is lost, however the cumulative meter data is still counting and shows the today and total values correctly once the internet link comes back.

Thanks - that does make things clearer.

Something of a pain - I've been writing scripts which use the API to estimate costs based on the 5 minute intervals and off-peak and peak tariffs - so this throws things out in annoying ways. Looks like I'm going to have to guess a few figures and fudge things for today.

I wonder if it's worth setting up my own logger that talks directly to the inverter, instead of going through the API.

michaelmcgoldrick Hi Michael.... just spotted your post. I noticed that my battery did NOT charge overnight. Did your chargeing start before the outage occured do you think?

MikeT Anyone else with similar omission of overnight battery charge?

Our battery charged just fine overnight - so looks like it was just data loss for us.

No data from midnight and started again 7.30am. How do you report to GivEnergy?

MikeT i Michael.... just spotted your post. I noticed that my battery did NOT charge overnight. Did your chargeing start before the outage occured do you think?

Looks like our data outage started on the dot of midnight last night, if I believe the inverter data on the dashboard, or 11pm if I believe the figures from API (BST issue?). So I don't think it was that.

straycat264

Thanks. Just to add that my battery was scheduled to start charging from 01:05 and stop an 04:00 but it did not do so. My "outage" started spot on midnight as well.

iBarryJ
With the cost or storage these days you should be able to log a lot more than that.

MikeT scheduled to charge from 2am to 8am and started at 2am finished by around 4am as its generally only half empty these days coming charge time

Ra

If I add up the number of missing data I've had since my system went live on the 11th Jan and don't even include the 9 days I lost due to a failed dongle.... It still adds up to 93hrs.

You may not think that's much downtime, but if our servers fell over that much when we have a 99.999% uptime contract... we'd be getting some massive refunds.

And just so people realise... a 99% uptime would allow for 87.6hrs of downtime... unacceptable in any industry that relies upon connectivity to clients. I've already passed that in less than 3 months.

We have a 99.999% uptime deal, that allows for but a few hrs downtime in an entire year.... and they've never dropped below that. They offered us a 99.9999% contract that's in the 'minutes' of downtime per year. But it was just too expensive.

Servers these days are ultra reliable... the kind of downtime we're seeing points to low quality kit, low quality service providers and/or low quality connections.... which is a sign of cost cutting and being unable to provide a basic service.

Similar issue here. Overnight timed charge 0045 to 0430. OK. But then (another) strange thing happened - at 0430 battery completely discharged itself, all exported to the grid so that by 0730 it was only 4% charge, down from 40% at midnight. Obviously don’t have anny timed discharge nor timed export configured. Anybody else see this?

Anomnomnomaly Absolutely agree. We run multiple servers in two UK data centres and even the non-critical boxes run with 99.999% uptime average. The "mission critical" stuff is configured with auto-failover and just doesn't fail - not cheap but then you have complete confidence. It seems like GE are running on a NAS box in someone's bedroom at times.

Anomnomnomaly How much do you pay for your 99.9999% uptime?

And how much do we pay Ge for the data we look at? Just musing................

sponev Even my NAS has 99.9999% uptime!

wilfleek last quote I did for 11 9s was £160k per year using Azure.

wilfleek was thinking the same the other day.

Some of the purchase price presumably goes towards our free lifetime (of the inverter) subscription to their cloud services!

wilfleek The service is included with the purchase of the equipment. If they can't provide that service, then they should have an open source api for people to use to grab the data themselves, and not restricted with proprietary software/access and restrictions to accessing data based on unreliable software/equipment.

wilfleek We don't pay them anything ongoing, their product model incorporates the remote management and user access. Not paying an ongoing access fee doesn't justify it being flaky though? But then they know that and hopefully are working on making it reliable.

wilfleek That's not my area... I might be able to find out. But I don;t deal with finances, I just know that the beancounters said no when the quote for the 4 9 service came in.

I run my own basic server at home too... that's been running on at least 100% service if I don't count the time I've had to turn the electric off to do work on the house, or swapping the system into a much larger case to expand storage. Or even moving home twice. If I do include those... downtime is still only in the 20-30hr range over the last 12 months.

I don't recall seeing the portal and its data touted as part of the deal when I researched the equipment? I view it as FOC, personally. It certainly did not figure in the ROI calculations I did.

Happy to be proved wrong - where did you see it as part of the deal/initial purchase price?

I fully expect GE to start charging for new features in the future. Personally I don't use the portal at all and I don't need it as I am able to do everything I need via Home Assistant but I can see that with time as the platform (and their hardware) matures they will probably introduce tiers with the free tier giving you basic functionality and the more advanced (useful) features being moved to the paid tier.

wilfleek Interesting. The deal clincher for me was the portal and array of data available, validated using the test account I was able to set up. Plus there are ample links to the supporting services from the GE website. It's clearly part of the product offer? Indeed knowing the SoC issue, if the portal and data hadn't been available, together with the promise of the SoC fix, I'd have gone elsewhere.