1 store + api + instance object done
2 agent + reconcile loop, isolation: container done
3 netdev + nft + dns → containers talk done
4 oci image store done
5 vmm/qemu → isolation: vm done
6 node join + routing → multi node done
7 isolation: microvm and sandbox done
8 serial console done
9 image catalog: disk, iso, kernel done
10 volumes done
11 api authentication done
12 volume snapshots done
13 fencing a partitioned node done
14 published ports + firewall done
15 usage metrics + load aware placement done
16 audit trail done
17 projects and per-project scoping done
18 off-node volume backup and restore done
19 token lifetimes + TLS on the API done
20 per-project quotas + read-only viewer role done
21 scheduled backups + retention done
22 backup encryption at rest done
23 encrypted volumes on the node done
24 backup and restore of encrypted volumes done
25 volume resize, placement groups, disk hot-plug done
26 ssh keys for vm instances done
27 load balancer across replicas done
28 health checked backends done
29 events: what the platform did on its own done
30 services: hold a replica count done
31 a balancer that follows a service done
32 cordon and drain a node done
33 bounded usage history done
34 webhooks: tell somebody an event happened done
35 backups on an S3 object store done
36 nodes write backups straight to the store done
37 paged instance listing done
38 paged volume, backup and snapshot listings done
39 operator labels on nodes done
40 placing a workload by node selector done (instances, services)
41 sealed environment injection done
42 sealed config file injection done
43 per caller api rate limiting done
44 resize an instance's cpu and memory done
45 tls termination on a balancer done
46 more than one network per instance done
47 ipv6 networks done
48 dual stack networks done
49 publishing either side of a dual stack instance done
50 firewalls covering every address of an instance done
51 env and config files on every isolation done
52 tls on a published port done
53 scheduled snapshots done
54 a rate limit per project done
55 rolling update of a service's template done
56 refusing a workload that names nothing done
57 replacing a replica the node gave up on done
58 reading what a workload printed done
59 work that finishes: jobs and schedules done
60 a service that holds its own replica count done
61 starting a container the node already holds done
62 people, sessions and who did it done
63 scaling on memory, not only cpu done
64 a container with a working /dev and /sys done
65 a container that can read its own limits done
66 calling an instance by its name done
67 paging the audit trail and the event log done
68 a snapshot that becomes a volume of its own done
69 taking a node out with a vm on it done
70 getting inside a running container done
71 one port serving many applications done
72 a start that can never work, said once done