Possibly a good time to remind people that the default value of jobs.<job_id>.timeout-minutes is 360 (minutes), meaning that your hanging job will cost $0.72 before it times out.
I learned this lesson the hard way recently. I'm on the 50k minutes plan and burned through my entire monthly allowance plus a $100 overage budget in about two weeks.
My mistake was a combination of triggering workflows on every push, using macOS runners (which I didn't realize had a 10x multiplier compared to Linux), and forgetting to set aggressive timeouts.
I'm sharing this because the support experience was actually the highlight. I opened a ticket expecting to just eat the cost, but they sent a detailed breakdown of my usage/mistakes the next day. Even though it was 100% my error (tho I used to think macOS runners were only 3x? True, did that change? anyway), they gave me a $50 coupon to offset the overage. Amidst the pricing discussions, I think it's worth noting that their support team is still very human and responsive.
I learned this lesson the hard way recently. I'm on the 50k minutes plan and burned through my entire monthly allowance plus a $100 overage budget in about two weeks.
My mistake was a combination of triggering workflows on every push, using macOS runners (which I didn't realize had a 10x multiplier compared to Linux), and forgetting to set aggressive timeouts.
I'm sharing this because the support experience was actually the highlight. I opened a ticket expecting to just eat the cost, but they sent a detailed breakdown of my usage/mistakes the next day. Even though it was 100% my error (tho I used to think macOS runners were only 3x? True, did that change? anyway), they gave me a $50 coupon to offset the overage. Amidst the pricing discussions, I think it's worth noting that their support team is still very human and responsive.