Cron PHP sessionclean hangs invoking oom-killer

This morning one of our Linodes (18.04.1 LTS (GNU/Linux 4.18.16-x86_64-linode118 x86_64)) crashed. From the syslog I can see that cron task PHP cleansession got stuck around 36 hours ago:

Dec 15 23:09:01 server systemd[1]: Starting Clean php session files...
Dec 15 23:09:01 server CRON[506]: (root) CMD (  [ -x /usr/lib/php/sessionclean ] && if [ ! -d /run/systemd/system ]; then /usr/lib/php/sessionclean; fi)
Dec 15 23:12:46 server kernel: [2456147.547283] INFO: task sessionclean:479 blocked for more than 120 seconds.
Dec 15 23:12:46 server kernel: [2456147.548953]       Not tainted 4.18.16-x86_64-linode118 #1
Dec 15 23:12:46 server kernel: [2456147.550155] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Dec 15 23:12:46 server kernel: [2456147.552176] sessionclean    D    0   479      1 0x80000002
Dec 15 23:12:46 server kernel: [2456147.552179] Call Trace:
Dec 15 23:12:46 server kernel: [2456147.552212]  ? __schedule+0x3be/0x8b0
Dec 15 23:12:46 server kernel: [2456147.552216]  ? uncharge_batch+0x2e9/0x3a0
Dec 15 23:12:46 server kernel: [2456147.552218]  schedule+0x3c/0x90
Dec 15 23:12:46 server kernel: [2456147.552220]  schedule_timeout+0x1f7/0x420
Dec 15 23:12:46 server kernel: [2456147.552225]  ? release_pages+0x2d2/0x3e0
Dec 15 23:12:46 server kernel: [2456147.552226]  wait_for_common+0xcd/0x190
Dec 15 23:12:46 server kernel: [2456147.552230]  ? wake_up_q+0x60/0x60
Dec 15 23:12:46 server kernel: [2456147.552235]  __wait_rcu_gp+0x11b/0x150
Dec 15 23:12:46 server kernel: [2456147.552238]  synchronize_rcu+0x7c/0x90
Dec 15 23:12:46 server kernel: [2456147.552241]  ? kfree_call_rcu+0x20/0x20
Dec 15 23:12:46 server kernel: [2456147.552243]  ? __bpf_trace_rcu_utilization+0x10/0x10
Dec 15 23:12:46 server kernel: [2456147.552246]  namespace_unlock+0x67/0x80
Dec 15 23:12:46 server kernel: [2456147.552249]  put_mnt_ns+0x1c/0x30
Dec 15 23:12:46 server kernel: [2456147.552253]  free_nsproxy+0x17/0x90
Dec 15 23:12:46 server kernel: [2456147.552256]  do_exit+0x3f0/0xc40
Dec 15 23:12:46 server kernel: [2456147.552259]  do_group_exit+0x39/0xb0
Dec 15 23:12:46 server kernel: [2456147.552261]  __x64_sys_exit_group+0x14/0x20
Dec 15 23:12:46 server kernel: [2456147.552265]  do_syscall_64+0x5b/0x190
Dec 15 23:12:46 server kernel: [2456147.552269]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
Dec 15 23:12:46 server kernel: [2456147.552272] RIP: 0033:0x7faf2120fe06
Dec 15 23:12:46 server kernel: [2456147.552272] Code: Bad RIP value.
Dec 15 23:12:46 server kernel: [2456147.552279] RSP: 002b:00007ffda594cd08 EFLAGS: 00000246 ORIG_RAX: 00000000000000e7
Dec 15 23:12:46 server kernel: [2456147.552281] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007faf2120fe06
Dec 15 23:12:46 server kernel: [2456147.552281] RDX: 0000000000000000 RSI: 000000000000003c RDI: 0000000000000000
Dec 15 23:12:46 server kernel: [2456147.552282] RBP: 0000556f16cd1a90 R08: 00000000000000e7 R09: ffffffffffffff80
Dec 15 23:12:46 server kernel: [2456147.552283] R10: 00007faf212c9cc0 R11: 0000000000000246 R12: 0000556f16cc0a20
Dec 15 23:12:46 server kernel: [2456147.552284] R13: 00007ffda594cff0 R14: 0000000000000000 R15: 0000000000000000

After that cron was failing every 5 minutes with the following error:

Dec 16 09:00:01 server kernel: [2491382.537416] cgroup: fork rejected by pids controller in /system.slice/cron.service
Dec 16 14:25:01 server cron[644]: (CRON) error (can't fork)

This morning we were notified by our backup service and monitor that SSH and websites were not responding anymore. CPU usage was at 100% for almost 4 hours:

Dec 17 05:05:04 server kernel: [2563686.476314] systemd invoked oom-killer: gfp_mask=0x6200ca(GFP_HIGHUSER_MOVABLE), nodemask=(null), order=0, oom_score_adj=0
Dec 17 05:05:05 server kernel: [2563686.476319] systemd cpuset=/ mems_allowed=0

Only solution was to reboot the system manually. Now everything is working again.

Any ideas why this was happening? We'd like to prevent those issues in the future.