zitadel/internal/database/cockroach/crdb.go
Silvan dab5d9e756
refactor(eventstore): move push logic to sql (#8816)
# Which Problems Are Solved

If many events are written to the same aggregate id it can happen that
zitadel [starts to retry the push
transaction](48ffc902cc/internal/eventstore/eventstore.go (L101))
because [the locking
behaviour](48ffc902cc/internal/eventstore/v3/sequence.go (L25))
during push does compute the wrong sequence because newly committed
events are not visible to the transaction. These events impact the
current sequence.

In cases with high command traffic on a single aggregate id this can
have severe impact on general performance of zitadel. Because many
connections of the `eventstore pusher` database pool are blocked by each
other.

# How the Problems Are Solved

To improve the performance this locking mechanism was removed and the
business logic of push is moved to sql functions which reduce network
traffic and can be analyzed by the database before the actual push. For
clients of the eventstore framework nothing changed.

# Additional Changes

- after a connection is established prefetches the newly added database
types
- `eventstore.BaseEvent` now returns the correct revision of the event

# Additional Context

- part of https://github.com/zitadel/zitadel/issues/8931

---------

Co-authored-by: Tim Möhlmann <tim+github@zitadel.com>
Co-authored-by: Livio Spring <livio.a@gmail.com>
Co-authored-by: Max Peintner <max@caos.ch>
Co-authored-by: Elio Bischof <elio@zitadel.com>
Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com>
Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com>
Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com>
Co-authored-by: Yxnt <Yxnt@users.noreply.github.com>
Co-authored-by: Stefan Benz <stefan@caos.ch>
Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com>
Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00

243 lines
6.1 KiB
Go

package cockroach
import (
"context"
"database/sql"
"fmt"
"strconv"
"strings"
"time"
"github.com/jackc/pgx/v5"
"github.com/jackc/pgx/v5/pgxpool"
"github.com/jackc/pgx/v5/stdlib"
"github.com/mitchellh/mapstructure"
"github.com/zitadel/logging"
"github.com/zitadel/zitadel/internal/api/authz"
"github.com/zitadel/zitadel/internal/database/dialect"
)
func init() {
config := new(Config)
dialect.Register(config, config, true)
}
const (
sslDisabledMode = "disable"
sslRequireMode = "require"
sslAllowMode = "allow"
sslPreferMode = "prefer"
)
type Config struct {
Host string
Port uint16
Database string
MaxOpenConns uint32
MaxIdleConns uint32
MaxConnLifetime time.Duration
MaxConnIdleTime time.Duration
User User
Admin AdminUser
// Additional options to be appended as options=<Options>
// The value will be taken as is. Multiple options are space separated.
Options string
}
func (c *Config) MatchName(name string) bool {
for _, key := range []string{"crdb", "cockroach"} {
if strings.TrimSpace(strings.ToLower(name)) == key {
return true
}
}
return false
}
func (_ *Config) Decode(configs []interface{}) (dialect.Connector, error) {
connector := new(Config)
decoder, err := mapstructure.NewDecoder(&mapstructure.DecoderConfig{
DecodeHook: mapstructure.StringToTimeDurationHookFunc(),
WeaklyTypedInput: true,
Result: connector,
})
if err != nil {
return nil, err
}
for _, config := range configs {
if err = decoder.Decode(config); err != nil {
return nil, err
}
}
return connector, nil
}
func (c *Config) Connect(useAdmin bool, pusherRatio, spoolerRatio float64, purpose dialect.DBPurpose) (*sql.DB, *pgxpool.Pool, error) {
dialect.RegisterAfterConnect(func(ctx context.Context, c *pgx.Conn) error {
// CockroachDB by default does not allow multiple modifications of the same table using ON CONFLICT
// This is needed to fill the fields table of the eventstore during eventstore.Push.
_, err := c.Exec(ctx, "SET enable_multiple_modifications_of_table = on")
return err
})
connConfig, err := dialect.NewConnectionConfig(c.MaxOpenConns, c.MaxIdleConns, pusherRatio, spoolerRatio, purpose)
if err != nil {
return nil, nil, err
}
config, err := pgxpool.ParseConfig(c.String(useAdmin, purpose.AppName()))
if err != nil {
return nil, nil, err
}
if len(connConfig.AfterConnect) > 0 {
config.AfterConnect = func(ctx context.Context, conn *pgx.Conn) error {
for _, f := range connConfig.AfterConnect {
if err := f(ctx, conn); err != nil {
return err
}
}
return nil
}
}
// For the pusher we set the app name with the instance ID
if purpose == dialect.DBPurposeEventPusher {
config.BeforeAcquire = func(ctx context.Context, conn *pgx.Conn) bool {
return setAppNameWithID(ctx, conn, purpose, authz.GetInstance(ctx).InstanceID())
}
config.AfterRelease = func(conn *pgx.Conn) bool {
ctx, cancel := context.WithTimeout(context.Background(), time.Second)
defer cancel()
return setAppNameWithID(ctx, conn, purpose, "IDLE")
}
}
if connConfig.MaxOpenConns != 0 {
config.MaxConns = int32(connConfig.MaxOpenConns)
}
config.MaxConnLifetime = c.MaxConnLifetime
config.MaxConnIdleTime = c.MaxConnIdleTime
pool, err := pgxpool.NewWithConfig(context.Background(), config)
if err != nil {
return nil, nil, err
}
if err := pool.Ping(context.Background()); err != nil {
return nil, nil, err
}
return stdlib.OpenDBFromPool(pool), pool, nil
}
func (c *Config) DatabaseName() string {
return c.Database
}
func (c *Config) Username() string {
return c.User.Username
}
func (c *Config) Password() string {
return c.User.Password
}
func (c *Config) Type() string {
return "cockroach"
}
func (c *Config) Timetravel(d time.Duration) string {
return ""
}
type User struct {
Username string
Password string
SSL SSL
}
type AdminUser struct {
// ExistingDatabase is the database to connect to before the ZITADEL database exists
ExistingDatabase string
User `mapstructure:",squash"`
}
type SSL struct {
// type of connection security
Mode string
// RootCert Path to the CA certificate
RootCert string
// Cert Path to the client certificate
Cert string
// Key Path to the client private key
Key string
}
func (c *Config) checkSSL(user User) {
if user.SSL.Mode == sslDisabledMode || user.SSL.Mode == "" {
user.SSL = SSL{Mode: sslDisabledMode}
return
}
if user.SSL.Mode == sslRequireMode || user.SSL.Mode == sslAllowMode || user.SSL.Mode == sslPreferMode {
return
}
if user.SSL.RootCert == "" {
logging.WithFields(
"cert set", user.SSL.Cert != "",
"key set", user.SSL.Key != "",
"rootCert set", user.SSL.RootCert != "",
).Fatal("at least ssl root cert has to be set")
}
}
func (c Config) String(useAdmin bool, appName string) string {
user := c.User
if useAdmin {
user = c.Admin.User
}
c.checkSSL(user)
fields := []string{
"host=" + c.Host,
"port=" + strconv.Itoa(int(c.Port)),
"user=" + user.Username,
"dbname=" + c.Database,
"application_name=" + appName,
"sslmode=" + user.SSL.Mode,
}
if c.Options != "" {
fields = append(fields, "options="+c.Options)
}
if !useAdmin {
fields = append(fields, "dbname="+c.Database)
} else if c.Admin.ExistingDatabase != "" {
fields = append(fields, "dbname="+c.Admin.ExistingDatabase)
}
if user.Password != "" {
fields = append(fields, "password="+user.Password)
}
if user.SSL.Mode != sslDisabledMode {
fields = append(fields, "sslrootcert="+user.SSL.RootCert)
if user.SSL.Cert != "" {
fields = append(fields, "sslcert="+user.SSL.Cert)
}
if user.SSL.Key != "" {
fields = append(fields, "sslkey="+user.SSL.Key)
}
}
return strings.Join(fields, " ")
}
func setAppNameWithID(ctx context.Context, conn *pgx.Conn, purpose dialect.DBPurpose, id string) bool {
// needs to be set like this because psql complains about parameters in the SET statement
query := fmt.Sprintf("SET application_name = '%s_%s'", purpose.AppName(), id)
_, err := conn.Exec(ctx, query)
logging.OnError(err).Warn("failed to set application name")
return err == nil
}