There must be a huge latency penalty with this setup, which would preclude some uses. But I’m wondering if you could use VS Code remote on a VM next to the k8s cluster and get almost the same dev experience without having the app see extra latency.
Hey, one of mirrord’s maintainers here. There’s of course extra latency as traffic has to go from your machine to the cluster and back, but as the common use case for mirrord isn’t performance-intensive to begin with (testing functionality of new code, or debugging existing issues in the cloud), this isn’t really a factor for most of its users.
But as you suggested, it is possible to use mirrord in tandem with a development machine within the cluster. We’re seeing this happen in organizations that already use a remote development environment solution like Gitpod.
Testing new code or debugging in a performance-intensive service is performance-intensive, right? Like, if I’m working on an authz service or database proxy and everything else expects it to respond in a millisecond, it’s going to be a problem. So it’s good there’s a way to get this experience (which sounds great) in that situation.
There must be a huge latency penalty with this setup, which would preclude some uses. But I’m wondering if you could use VS Code remote on a VM next to the k8s cluster and get almost the same dev experience without having the app see extra latency.
Hey, one of mirrord’s maintainers here. There’s of course extra latency as traffic has to go from your machine to the cluster and back, but as the common use case for mirrord isn’t performance-intensive to begin with (testing functionality of new code, or debugging existing issues in the cloud), this isn’t really a factor for most of its users.
But as you suggested, it is possible to use mirrord in tandem with a development machine within the cluster. We’re seeing this happen in organizations that already use a remote development environment solution like Gitpod.
Testing new code or debugging in a performance-intensive service is performance-intensive, right? Like, if I’m working on an authz service or database proxy and everything else expects it to respond in a millisecond, it’s going to be a problem. So it’s good there’s a way to get this experience (which sounds great) in that situation.
Is this a common opinion? Curious to hear about certifications devs took and didn’t later think their time would’ve been better spent otherwise.
Wondering how
mirrorddiffers from https://www.telepresence.io? Can anyone shed light?If you read the article until the end, you will find an answer.
Hey, in addition to the differences listed in the article (centered on concurrent usage of the shared environment), there are the differences listed here: https://mirrord.dev/docs/overview/faq/#how-is-mirrord-different-from-telepresence
tl;dr: