Is this normal, or not good - in terms of server monitoring

bikegremlin · May 2020

I had a server related problem (redis cache was not working properly).

Contacted tech. support, which blamed WordPress plugins.
Figured out on my own - first disabled one plugin at a time to find out LiteSpeed is the culprit. Then, thanks to how well made LiteSpeed plugin is - it showed what the error was after some digging through the menus.
Only after providing that info to the tech. support did they reset redis and it all worked.

In detail:
https://io.bikegremlin.com/12585/hosting-sajt-problemi/#p3

Now - my dilemma is: are things like this the job of whoever manages the server - to check whether all the services work properly (especially after updates)? Or is it just one of those things that can't be avoided.
Asking because I have no idea how it looks from "that side".

Not blaming the tech. support for this - the original info I had provided was really not enough to work with in my opinion.

Mr_Tom · May 2020

Depends on the service.

If it's an unmanaged VPS, then it's your issue. If it's managed it's theirs.

If it's on a shared hosting platform, it's probably their issue but they will often assume its your issue if no one else is seeing issues.

whoami · May 2020

I guess it depends on how severe the problem was. If it was causing your site to throw 500 errors continually and it's a managed service, I'd be disappointed they weren't proactive.

But if it's like "oh this one little feature buried under a search menu doesn't work and it's an unmanaged shared hosting account" then yeah, their response is ok.

mikho · May 2020

I would say it depends on the contract.
If it was fully managed, meaning you don’t have root, and they have done the installation.

Then it’s there issue to fix. If you have root, then you must check the contract to see what parts they manage.

Most cases goes by, If you have installed it, you take care of it.
Unless some 3rd party has crashed it.

bikegremlin · May 2020

Can't edit the original post:

Shared (reseller) hosting environment - important, yet forgot to mention that.

Problem manifested in error_log being filled with errors (that's how I had noticed it).
And, redis object cache not working - but sites loaded fine.

They did fix it. After I had pinpointed the problem cause. The question is whether monitoring such stuff (and fixing it before it gets reported) is provider's job, or not.

Not looking to blame anyone, but it will affect my decision on whether a hosting provider is good/reliable, or to look further (if such things keep happening).

Mr_Tom · May 2020

I think if they're offering redis across the server for all shared accounts then perhaps it should be monitored, but it depends on price.

If it's cheap hosting, they might just monitor the server/port 80/etc directly and see addons like redis as low priority and just wait for users to report issues.

bikegremlin · May 2020

@Mr_Tom said:
I think if they're offering redis across the server for all shared accounts then perhaps it should be monitored, but it depends on price.

If it's cheap hosting, they might just monitor the server/port 80/etc directly and see addons like redis as low priority and just wait for users to report issues.

I'd put price out of the equation in this case - when the aim is to determine good/excellent/poor.

Of course, I understand that you can't expect excellent quality at budget price. But it's good to know when that is the case.

Mr_Tom · May 2020

It probably should be monitored if it's part of the package, but due to cost they might just be reliant on clients reporting issues with it - or their monitoring failed.

I would say it's probably not uncommon for them to ask you to make sure it's not your site that has the issue before they investigate. I've had that plenty of times before when dealing with managed servers too unfortunately.

Janevski · June 2020

Depends on what you both agreed upon.
Usually fully managed services come at a very high price.

I doubt that someone using shared hosting is paying for fully managed service.
Usually things like, own server or at least a decent VDS, go without saying.

Only difference are niche services which are specialized in providing only one of a kind shared hosting, they do manage the techincal part of the service, however you're responsible for managing the content, for example.

InceptionHosting · June 2020

I personally put it down to one of those things that happens, if it turns out to be one of 10 things that 'just happened' in a month then maybe look to use another host.

With shared hosting especially there are so many combinations of things that may cause an issue under the right (wrong?) circumstances that unless it is a premium service that pays enough to have a dedicated person regularly monitoring every log then on occasion things will go wrong.

Sounds like its a cheap or fairly cheap service to me though whereby it is more cost-effective (for you and them) o simply fix on fail which is an incredibly common model in the tech world.

bikegremlin · June 2020

@AnthonySmith said:
I personally put it down to one of those things that happens, if it turns out to be one of 10 things that 'just happened' in a month then maybe look to use another host.

With shared hosting especially there are so many combinations of things that may cause an issue under the right (wrong?) circumstances that unless it is a premium service that pays enough to have a dedicated person regularly monitoring every log then on occasion things will go wrong.

Sounds like its a cheap or fairly cheap service to me though whereby it is more cost-effective (for you and them) o simply fix on fail which is an incredibly common model in the tech world.

This sums up my first thoughts on it.

Being a lazy bastard, I'm curious if it is possible to have some monitoring set up, to be notified if a service is not working. At least as a server admin - if not as a customer (no root access etc). Primarily out of curiosity (not planning to run even an unmanaged VPS any time soon).

Is this normal, or not good - in terms of server monitoring

Comments