| ▲ | asa400 21 hours ago |
| > Making GitHub actions runners more reliable, stable and economical would hurt Microsoft. Can you explain this one a bit? I know some folks who would absolutely increase their spend if Actions runners were better. As it stands I've seen folks trying to move off of them. |
|
| ▲ | gnopgnip 15 hours ago | parent | next [-] |
| If they were twice as efficient, you would only need to run half as many and spend less. |
|
| ▲ | heliumtera 21 hours ago | parent | prev [-] |
| Gh actions runners had a dubious implementation of sleep that would cause runners to hang on 100% usage for weeks/months. A simple fix was proposed and neglected for 10 years.
This discussion resurfaced recently with zig abandoning GitHub entirely and criticizing this specific issue.
A fix was them merged following an announcement that self hosted runners will now be charged by the minute. Of course this two facts are totally independent but yeah, yeah, sure. |
| |
| ▲ | staticassertion 20 hours ago | parent [-] | | How does this support your point? If we're saying "they fixed it because there was pressure to"... okay? That's the parents point - tons of people are going to move off over bad performance, and Github was incentivized to fix it when people started moving off. If Github's incentive was to keep it slow... we wouldn't have seen exactly what you're describing. | | |
| ▲ | heliumtera 20 hours ago | parent [-] | | The fix already existed and was neglected for a decade. It was a 3 lines of bash code.
The big would commonly make a runner hang forever unnoticed, on a platform that charged by the minute.
One minute hanging, was one minute charged.
The fix that would drop considerably the amount of total minutes charged was immediately followed by charging self hosted runners by the minute. >GitHub incentive was to make it fast. They charge by the minute. The faster it completes the action, less money. Runner go fast pocket go low | | |
|
|
