Bad write log efficiency, strange broken relationships

Question

I recently switched galera synchronization method from rsync to mariabackup which greatly reduced CPU usage on our DB servers and also increased the innodb_log_file_size to 2G from about 600M or so. Since then (might be a coincidence) I started getting a lot of aborted connection errors and bad log write efficiency. I initially thought it had something to do with wait_timeout (which we set to 5 since our application has to be very fast ) and raised it to 60 as well as max_allowed_packet to 2G. But even after that the errors still remained. And AlertManager raises host crash errors almost hourly.

[--] Skipped version check for MySQLTuner script
[OK] Logged in using credentials passed on the command line
[OK] Currently running supported MySQL version 10.4.14-MariaDB-1:10.4.14+maria~focal-log
[OK] Operating on 64-bit architecture
 
-------- Log file Recommendations ------------------------------------------------------------------
[!!] Log file  doesn't exist
 
-------- Storage Engine Statistics -----------------------------------------------------------------
[--] Status: +Aria +CSV +InnoDB +MEMORY +MRG_MyISAM +MyISAM +PERFORMANCE_SCHEMA +SEQUENCE 
[--] Data in InnoDB tables: 16.3G (Tables: 150)
[OK] Total fragmented tables: 0
 
-------- Analysis Performance Metrics --------------------------------------------------------------
[--] innodb_stats_on_metadata: OFF
[OK] No stat updates during querying INFORMATION_SCHEMA.
 
-------- Security Recommendations ------------------------------------------------------------------
[OK] There are no anonymous accounts for any database users
[!!] User 'mariadb.sys@localhost' has no password set.
[!!] User 'developer@%' does not specify hostname restrictions.
[--] There are 620 basic passwords in the list.
 
-------- CVE Security Recommendations --------------------------------------------------------------
[OK] NO SECURITY CVE FOUND FOR YOUR VERSION
 
-------- Performance Metrics -----------------------------------------------------------------------
[--] Up for: 4h 56m 41s (1M q [64.284 qps], 33K conn, TX: 758M, RX: 166M)
[--] Reads / Writes: 40% / 60%
[--] Binary logging is enabled (GTID MODE: OFF)
[--] Physical Memory     : 54.9G
[--] Max MySQL memory    : 46.1G
[--] Other process memory: 0B
[--] Total buffers: 19.2G global + 26.9M per thread (1000 max threads)
[--] P_S Max memory usage: 623M
[--] Galera GCache Max memory usage: 128M
[OK] Maximum reached memory usage: 20.7G (37.77% of installed RAM)
[OK] Maximum possible memory usage: 46.1G (83.90% of installed RAM)
[OK] Overall possible memory usage with other process is compatible with memory available
[OK] Slow queries: 4% (55K/1M)
[OK] Highest usage of available connections: 3% (36/1000)
[!!] Aborted connections: 10.61%  (3560/33563)
[OK] Query cache is disabled by default due to mutex contention on multiprocessor machines.
[OK] Sorts requiring temporary tables: 0% (0 temp sorts / 3 sorts)
[OK] No joins without indexes
[OK] Temporary tables created on disk: 12% (13 on disk / 107 total)
[OK] Thread cache hit rate: 99% (69 created / 33K connections)
[OK] Table cache hit rate: 97% (290 open / 296 opened)
[OK] table_definition_cache(400) is upper than number of tables(313)
[OK] Open file limit used: 0% (58/16K)
[OK] Table locks acquired immediately: 100% (686 immediate / 686 locks)
[OK] Binlog cache memory access: 100.00% (586009 Memory / 586009 Total)
 
-------- Performance schema ------------------------------------------------------------------------
[--] Memory used by P_S: 623.7M
[--] Sys schema isn't installed.
 
-------- ThreadPool Metrics ------------------------------------------------------------------------
[--] ThreadPool stat is enabled.
[--] Thread Pool Size: 14 thread(s).
[--] Using default value is good enough for your version (10.4.14-MariaDB-1:10.4.14+maria~focal-log)
 
-------- MyISAM Metrics ----------------------------------------------------------------------------
[!!] Key buffer used: 18.3% (6M used / 33M cache)
[!!] Cannot calculate MyISAM index size - re-run script as root user
 
-------- InnoDB Metrics ----------------------------------------------------------------------------
[--] InnoDB is enabled.
[--] InnoDB Thread Concurrency: 0
[OK] InnoDB File per table is activated
[OK] InnoDB buffer pool / data size: 19.0G/16.3G
[OK] Ratio InnoDB log file size / InnoDB Buffer pool size: 2.0G * 2/19.0G should be equal to 25%
[OK] InnoDB buffer pool instances: 19
[--] Number of InnoDB Buffer Pool Chunk : 152 for 19 Buffer Pool Instance(s)
[OK] Innodb_buffer_pool_size aligned with Innodb_buffer_pool_chunk_size & Innodb_buffer_pool_instances
[OK] InnoDB Read buffer efficiency: 100.00% (12044117752 hits/ 12044442245 total)
[!!] InnoDB Write Log efficiency: 87.27% (874238 hits/ 1001803 total)
[OK] InnoDB log waits: 0.00% (0 waits / 127565 writes)
 
-------- AriaDB Metrics ----------------------------------------------------------------------------
[--] AriaDB is enabled.
[OK] Aria pagecache size / total Aria indexes: 128.0M/312.0K
[OK] Aria pagecache hit rate: 99.5% (2K cached / 14 reads)
 
-------- TokuDB Metrics ----------------------------------------------------------------------------
[--] TokuDB is disabled.
 
-------- XtraDB Metrics ----------------------------------------------------------------------------
[--] XtraDB is disabled.
 
-------- Galera Metrics ----------------------------------------------------------------------------
[--] Galera is enabled.
[--] GCache is using 0B
[--] CPU core detected  : 14

[--] wsrep_slave_threads: 32
[OK] wsrep_slave_threads is equal to 2, 3 or 4 times number of CPU(s)
[OK] gcs.fc_limit should be equal to 5 * wsrep_slave_threads
[--] wsrep parallel slave can cause frequent inconsistency crash.
[OK] gcs.fc_limit is equal to 5 * wsrep_slave_threads
[OK] gcs.fc_factor is equal to 0.8
[OK] Flow control fraction seems to be OK (wsrep_flow_control_paused<=0.02)
[OK] All tables are InnoDB tables
[OK] Binlog format is in ROW mode.
[OK] InnoDB flush log at each commit is disabled for Galera.
[--] Read consistency mode :OFF
[OK] Galera WsREP is enabled.
[OK] Galera Cluster address is defined: gcomm://10.10.5.151,10.10.5.152,10.10.5.153
[--] There are 3 nodes in wsrep_cluster_address
[OK] There are 3 nodes in wsrep_cluster_size.
[OK] All cluster nodes detected.
[OK] Galera Cluster name is defined: fraudsniper_cluster
[OK] Galera Node name is defined: node1
[!!] Galera Notify command is not defined.
[OK] SST Method is based on xtrabackup.
[OK] TOI is default mode for upgrade.
[--] Max WsRep message : 2.0G
[OK] Node is connected
[OK] Node is ready
[--] Cluster status :Primary
[OK] Galera cluster is consistent and ready for operations
[OK] Node and whole cluster at the same level: d1108e1b-bb16-11ea-b339-c7a9c5da38c9
[OK] Node is synced with whole cluster.
[OK] There is no certification failures detected.
 
-------- Replication Metrics -----------------------------------------------------------------------
[--] Galera Synchronous replication: YES
[--] No replication slave(s) for this server.
[--] Binlog format: ROW
[--] XA support enabled: ON
[--] Semi synchronous replication Master: OFF
[--] Semi synchronous replication Slave: OFF
[--] This is a standalone server
 
-------- Recommendations ---------------------------------------------------------------------------
General recommendations:
    Set up a Secure Password for mariadb.sys@localhost user: SET PASSWORD FOR 'mariadb.sys'@'SpecificDNSorIp' = PASSWORD('secure_password');
    Restrict Host for 'developer'@% to developer@SpecificDNSorIp
    UPDATE mysql.user SET host ='SpecificDNSorIp' WHERE user='developer' AND host ='%'; FLUSH PRIVILEGES;
    MySQL was started within the last 24 hours - recommendations may be inaccurate
    Reduce or eliminate unclosed connections and network issues
    Consider installing Sys schema from https://github.com/mysql/mysql-sys for MySQL
    Consider installing Sys schema from https://github.com/FromDual/mariadb-sys for MariaDB
Variables to adjust:
    Set wsrep_slave_threads to 1 in case of HA_ERR_FOUND_DUPP_KEY crash on slave
    set up parameter wsrep_notify_cmd to be notify

I'm really at a loss on how to solve the issue. I've tried reducing the log size to 1G, doesn't seem to help either. I've also tried using the command:

SELECT * FROM INFORMATION_SCHEMA.PROCESSLIST where time>1 and command<>"Sleep";

To try debugging slow queries but it's always an empty set yet the tuner reports over 5k slow queries (the slow query time is set to 5s).