DigitalOcean Connectivity Issues (opens in new tab)

(status.digitalocean.com)

29 pointst3ra8y ago25 comments

25 comments

The HN title i clicked said 5 hours, is that true? Holy cow.

edit: I guess it is, the Tweet is from 12AM (PST) and it's currently 6AM, yet their status page shows mostly down regions. Though hard to say exactly what the red x means.

mdekkers8y ago

Not sure why the title was changed. They have been down for about 6 hours now, and people globally are reporting the inability to connect to their droplets. Their status updates are meaningless.

t3raOP8y ago

Yep post titled got edited (maybe mods ?)

ajross8y ago

Indeed, my droplet in SFO2 is still down, apparently for the last seven and a half hours at this point.

justaaron8y ago

my droplets are all fine, and were running the whole time. the issue seems to be connectivity related, as i can only ssh into most of them, and they've been coming back online in groups...

chrissnell8y ago

They have updated the status page with more information. Apparently, they had a major power outage that took out multiple redundant power sources.

ajross8y ago

At exactly midnight local time at the data center? Something doesn't add up. This is going to be one fun post mortem...

mdekkers8y ago

And the updates from their status channels are a joke.

tehbeard8y ago

Less of a joke than AWS green ticks.

mdekkers8y ago

nah, I think they are about the same.

mattcoles8y ago

Prod database deleted again? ;)

Bit of a shame to happen so close to the last one.

rashthedude8y ago

Lol

piniyini8y ago

Who to use as a backup to DO, is Linode a good choice?

mdekkers8y ago

I tried VULTR, but they also suck. Linode looks interesting, but they don't appear to offer private networking.

I was spending hundreds of dollars a month on DO, VULTR and a few others. Last month I rented 3x 16 core 64GB bare metal servers, stuck Ceph on, and am now standing up Mesosphere on KVM. Cheaper, faster, and not more of an operational risk over using the cloud-clowns.

lykron8y ago

Linode offers a shared private network per datacenter.

1 more reply

lykron8y ago

Linode, except for DDoSs, has been incredibly solid.

sitepodmatt8y ago

Aside from the DDOS incidents Linode is generally very solid in connectivity department, but their communication skills when things go wrong leaves a lot of be desired - often their status page is 20+ minutes behind and has omissions. The way they handled communication on security issues was equally poor.

2 more replies

piniyini8y ago

Thanks, I was leaning towards them myself. Will sign up.

1 more reply

joshmn8y ago

Vultr, OVH, Scaleway all come to mind for hourlies.

wkoszek8y ago

My droplets are still down 5 hours later.

j / k navigate · click thread line to collapse

25 comments

notheguyouthink8y ago

The HN title i clicked said 5 hours, is that true? Holy cow.

edit: I guess it is, the Tweet is from 12AM (PST) and it's currently 6AM, yet their status page shows mostly down regions. Though hard to say exactly what the red x means.

mdekkers8y ago

Not sure why the title was changed. They have been down for about 6 hours now, and people globally are reporting the inability to connect to their droplets. Their status updates are meaningless.

t3raOP8y ago

Yep post titled got edited (maybe mods ?)

ajross8y ago

Indeed, my droplet in SFO2 is still down, apparently for the last seven and a half hours at this point.

justaaron8y ago

my droplets are all fine, and were running the whole time. the issue seems to be connectivity related, as i can only ssh into most of them, and they've been coming back online in groups...

chrissnell8y ago

They have updated the status page with more information. Apparently, they had a major power outage that took out multiple redundant power sources.

ajross8y ago

At exactly midnight local time at the data center? Something doesn't add up. This is going to be one fun post mortem...

mdekkers8y ago

And the updates from their status channels are a joke.

tehbeard8y ago

Less of a joke than AWS green ticks.

mdekkers8y ago

nah, I think they are about the same.

mattcoles8y ago

Prod database deleted again? ;)